An intelligent control method and system for mechanical harvesting of rapeseed based on online monitoring

Abstract

The application discloses a rapeseed mechanized harvesting intelligent control method based on online monitoring, and relates to the field of intelligent control, which comprises the following steps: acquiring rapeseed growth correlation information of a target rapeseed field, predicting a harvesting time node based on the rapeseed growth correlation information; acquiring rapeseed remote sensing information at the predicted harvesting time node; dividing a plurality of grid rapeseed fields based on the rapeseed remote sensing information; calculating the ripeness of rapeseed in all grid rapeseed fields; screening rapeseed test cutting areas from all grid rapeseed fields; counting the test cutting loss rate of the rapeseed test cutting areas, and analyzing the total loss of rapeseed in the target rapeseed field; judging whether the predicted harvesting time node is an optimal harvesting node; and generating an optimal parameter interval of a rapeseed harvesting device. The application can effectively reduce the harvesting loss rate of a rapeseed field with a large area.

Description

Technical Field

[0001] This application relates to the field of intelligent control, and in particular to an intelligent control method and system for mechanized rapeseed harvesting based on online monitoring. Background Technology

[0002] Rapeseed, as an economic crop, is a fundamental source of edible oil. With the advancement of agricultural mechanization, the planting and harvesting areas of rapeseed have been expanding year by year, and the resulting loss rate problem has become increasingly prominent. This is because during the harvesting of rapeseed using mechanized equipment, losses are inevitable due to factors such as the performance of the harvesting equipment, harvesting weather, and the maturity of the rapeseed. This is especially true in areas with large rapeseed planting areas, where the total loss is even higher, causing serious economic losses to agricultural workers engaged in rapeseed cultivation. Therefore, a method is needed to minimize the rapeseed harvesting loss rate.

[0003] Existing methods for reducing loss rates primarily involve first selecting appropriate harvesting equipment and methods based on the rapeseed variety and area, then analyzing the overall maturity of the rapeseed field, and adjusting the equipment parameters accordingly to ensure the equipment harvests the rapeseed at suitable parameters. This method can reduce rapeseed harvesting losses to some extent. However, for large rapeseed fields, the maturity of rapeseed may vary in different areas, leading to different suitable equipment parameters for each area. Furthermore, the large size of the rapeseed field makes it difficult to comprehensively assess rapeseed maturity, potentially resulting in significant errors in maturity assessment. This could lead to inappropriate equipment parameter selection, making it difficult to further reduce rapeseed harvesting losses. Summary of the Invention

[0004] This application provides an intelligent control method and system for mechanized rapeseed harvesting based on online monitoring, which addresses the problem that existing technologies are insufficient to further reduce the harvesting loss rate in large rapeseed fields.

[0005] To achieve the above objectives, the embodiments of this application adopt the following technical solutions:

[0006] Firstly, a method for intelligent control of mechanized rapeseed harvesting based on online monitoring is provided, the method comprising:

[0007] Obtain rapeseed growth association information of the target rapeseed field, including soil information of the target rapeseed field, meteorological information of the location of the target rapeseed field, and sowing time and rapeseed variety information of the rapeseed planted in the target rapeseed field;

[0008] The rapeseed variety information, meteorological information, and soil information are input into a pre-constructed rapeseed growth model to predict the rapeseed growth cycle. Based on the rapeseed growth cycle prediction results, the predicted maturity time period of the target rapeseed field is determined.

[0009] The predicted harvest time node of the target rapeseed field is determined based on the sowing time information and the predicted maturity time period;

[0010] At the predicted harvest time, remote sensing information of rapeseed in the target rapeseed field is collected by a remote sensing drone;

[0011] Based on the remote sensing information of rapeseed, the size information of the target rapeseed field is obtained, and the target rapeseed field is divided into grids according to the size information of the rapeseed field to obtain multiple grid rapeseed fields;

[0012] Based on the rapeseed remote sensing information, feature analysis is performed on all the grid rapeseed fields to obtain rapeseed remote sensing features, and the rapeseed maturity of all the grid rapeseed fields is calculated based on the rapeseed remote sensing features.

[0013] Based on the rapeseed maturity, a benchmark grid rapeseed field is selected from all the grid rapeseed fields. Taking the benchmark grid rapeseed field as the center and the maximization of the rapeseed maturity difference as the selection objective, the adjacent grid rapeseed fields of the benchmark grid rapeseed field are iteratively diffused and screened until the number of selected rapeseed fields is greater than or equal to a preset number threshold, and the rapeseed trial harvesting area is obtained.

[0014] The rapeseed harvesting equipment is controlled to perform rapeseed trial harvesting tasks in the rapeseed trial harvesting area. During the execution of the rapeseed trial harvesting tasks, the rapeseed harvesting equipment is used to count the trial harvesting loss rate in the rapeseed trial harvesting area. Based on the trial harvesting loss rate and using correlation analysis, the total rapeseed loss of the target rapeseed field is analyzed.

[0015] The rapeseed maturity node of the target rapeseed field is determined by combining the rapeseed remote sensing features and the total rapeseed loss, and the predicted harvest time node is determined as the optimal harvest time node based on the rapeseed maturity node.

[0016] If the predicted harvest time node is the optimal harvest node, then the optimal parameter range of the rapeseed harvesting equipment is generated by combining the rapeseed maturity and the trial harvesting loss rate and using a parameter optimization algorithm, and the rapeseed harvesting equipment is controlled to harvest rapeseed in the target rapeseed field according to the optimal parameters.

[0017] Optionally, the step of performing feature analysis on all the grid rapeseed fields based on the rapeseed remote sensing information to obtain rapeseed remote sensing features, and calculating the rapeseed maturity of all the grid rapeseed fields based on the rapeseed remote sensing features, includes the following steps:

[0018] Preprocess the remote sensing information of all the rapeseed fields in the grid;

[0019] For any of the grid rapeseed fields, extract the reflectance spectrum from the preprocessed rapeseed remote sensing information;

[0020] The spectral reflectance value is calculated based on the reflected spectrum, and the spectral reflectance value is used to extract features to obtain the spectral features of rapeseed.

[0021] Principal component transformation is performed on the rapeseed remote sensing information to obtain the spatial color features of the rapeseed remote sensing information;

[0022] Clustering algorithms are used to classify the spatial color features to obtain the rapeseed color features of the rapeseed remote sensing information;

[0023] The rapeseed spectral features and rapeseed color features are integrated into the rapeseed remote sensing features of the grid rapeseed field;

[0024] A rapeseed maturity prediction model is constructed based on a neural network model, and all the remote sensing features of rapeseed are input into the trained rapeseed maturity prediction model. The rapeseed maturity prediction model outputs the rapeseed maturity of all the grid rapeseed fields.

[0025] Optionally, the step of constructing a rapeseed maturity prediction model based on a neural network model, inputting all the rapeseed remote sensing features into the trained rapeseed maturity prediction model, and outputting the rapeseed maturity of all the grid rapeseed fields through the rapeseed maturity prediction model includes the following steps:

[0026] A rapeseed maturity prediction model was constructed based on a BP neural network model.

[0027] The rapeseed maturity prediction model is trained using a pre-constructed rapeseed training set;

[0028] For any of the grid rapeseed fields, the rapeseed spectral characteristics and rapeseed color characteristics of the grid rapeseed fields are sequentially input into the rapeseed maturity prediction model to predict maturity, and a first maturity prediction value and a second maturity prediction value are obtained.

[0029] The first maturity prediction value and the second maturity prediction value are filtered and corrected using Kalman filtering;

[0030] The first maturity prediction value and the second maturity prediction value after filtering correction are fused to obtain the rapeseed maturity of all the grid rapeseed fields.

[0031] Optionally, the step of selecting several target grid rapeseed fields from all the grid rapeseed fields as rapeseed trial harvesting areas based on the rapeseed maturity includes the following steps:

[0032] Based on the rapeseed maturity and using the threshold method, the maturity level of all the grid rapeseed fields is divided into different grades to obtain the rapeseed maturity level of all the grid rapeseed fields.

[0033] If the rapeseed maturity level is the maximum rapeseed maturity level or the rapeseed maturity level is the minimum rapeseed maturity level, then all the grid rapeseed fields corresponding to the maximum rapeseed maturity level and the minimum rapeseed maturity level are marked as trial rapeseed fields respectively.

[0034] For any of the aforementioned trial-harvested rapeseed fields, the trial-harvested rapeseed field marking step is repeatedly performed based on the rapeseed maturity level and centered on the trial-harvested rapeseed field, until the number of trial-harvested rapeseed fields is greater than or equal to a preset threshold. The trial-harvested rapeseed field marking step is as follows:

[0035] Calculate the maturity grade difference between the adjacent rapeseed fields and the test rapeseed field, and record the adjacent rapeseed fields whose maturity grade difference is greater than or equal to a preset difference threshold as test rapeseed fields;

[0036] By integrating all the adjacent rapeseed fields that were previously tested, multiple initial test areas were obtained;

[0037] Construct a target maturity level matrix for the target rapeseed field based on the rapeseed maturity levels of all the grid rapeseed fields;

[0038] For any of the initial trial harvesting areas, a regional maturity level matrix for the initial trial harvesting area is constructed based on the rapeseed maturity level.

[0039] The similarity between the target maturity level matrix and all the regional maturity level matrices is calculated based on the similarity calculation algorithm to obtain multiple matrix similarities;

[0040] The initial trial cutting area corresponding to the highest data similarity is taken as the rapeseed trial cutting area.

[0041] Optionally, the controlled rapeseed harvesting equipment performs a rapeseed trial harvesting task in the rapeseed trial harvesting area. During the execution of the rapeseed trial harvesting task, the rapeseed harvesting equipment statistically analyzes the trial harvesting loss rate of the rapeseed trial harvesting area. Based on the trial harvesting loss rate and using correlation analysis, the total rapeseed loss of the target rapeseed field is analyzed, including the following steps:

[0042] When the rapeseed harvesting equipment reaches any of the grid rapeseed fields in the rapeseed trial harvesting area, the equipment parameters of the rapeseed harvesting equipment are updated based on the rapeseed maturity of the grid rapeseed field to obtain the harvesting equipment parameters;

[0043] The rapeseed harvesting equipment is controlled to perform rapeseed trial harvesting tasks in the rapeseed trial harvesting area according to the harvesting equipment parameters, and rapeseed trial harvesting data of all the grid rapeseed fields in the rapeseed trial harvesting area are obtained;

[0044] The trial harvesting loss rate of all grid rapeseed fields in the rapeseed trial harvesting area was calculated based on the rapeseed trial harvesting data.

[0045] A correlation analysis was performed on the trial harvest loss rate and the rapeseed maturity to obtain the correlation analysis results.

[0046] The rapeseed loss rate of all grid rapeseed fields, excluding the rapeseed trial harvesting area, was calculated based on the correlation analysis results.

[0047] The average rapeseed loss rate of the target rapeseed field is calculated by combining all the trial harvest loss rates and the rapeseed loss rate.

[0048] The total rapeseed loss of the target rapeseed field is calculated based on the average rapeseed loss rate and the pre-obtained estimated rapeseed yield.

[0049] Optionally, updating the equipment parameters of the rapeseed harvesting equipment based on the rapeseed maturity in the grid rapeseed field to obtain the harvesting equipment parameters includes the following steps:

[0050] Image information of rapeseed in the grid rapeseed field is obtained using an image acquisition device pre-installed outside the rapeseed harvesting equipment;

[0051] Preprocess the rapeseed image information;

[0052] The preprocessed rapeseed image information is binarized using a threshold segmentation method to obtain the target rapeseed image.

[0053] Obtain the device parameters of the image acquisition device, and calculate the rapeseed height information of the rapeseed test cutting area based on the device parameters and the target rapeseed image;

[0054] The rapeseed harvesting equipment parameters are updated by combining the rapeseed height information and the rapeseed maturity corresponding to the grid rapeseed field, thus obtaining the harvesting equipment parameters.

[0055] Optionally, the step of determining the rapeseed maturity node of the target rapeseed field by combining the rapeseed remote sensing features and the total rapeseed loss, and determining whether the predicted harvest time node is the optimal harvest node based on the rapeseed maturity node, includes the following steps:

[0056] If the total loss of rapeseed is less than or equal to a preset loss threshold, then the predicted harvest time node is determined to be the optimal harvest node.

[0057] If the total loss of rapeseed is greater than the loss threshold, then the rapeseed maturity node of the target rapeseed field is predicted by combining the rapeseed image information and the rapeseed remote sensing features.

[0058] If the rapeseed maturity point is later than the predicted harvest time point, then the predicted harvest time point is determined not to be the optimal harvest time point;

[0059] If the rapeseed maturity point is not later than the predicted harvest time point, then the predicted harvest time point is determined to be the optimal harvest time point.

[0060] Optionally, the step of predicting the rapeseed maturity node of the target rapeseed field by combining the rapeseed image information and the rapeseed remote sensing features includes the following steps:

[0061] A rapeseed maturity prediction model was constructed based on a neural network model and a self-attention mechanism, and the model was trained using a pre-constructed training set.

[0062] The root and stem regions of the rapeseed image information are identified using a target detection algorithm. Based on the root and stem region identification results, the rapeseed image information is segmented to obtain the rapeseed root and stem image in the rapeseed image information.

[0063] Clustering algorithms were used to extract the root and stem color features of the rapeseed root and stem images;

[0064] The root and stem color features, rapeseed height information, and rapeseed remote sensing features are input into the trained rapeseed maturity prediction model, and the predicted rapeseed maturity period of the target rapeseed field is output through the rapeseed maturity prediction model.

[0065] The predicted harvest time is corrected based on the predicted rapeseed maturity date to obtain the rapeseed maturity date of the target rapeseed field.

[0066] Optionally, the step of combining the rapeseed maturity and the trial harvesting loss rate and using a parameter optimization algorithm to generate the optimal parameter range for the rapeseed harvesting equipment includes the following steps:

[0067] All grid rapeseed fields in the rapeseed trial harvesting area whose rapeseed loss rate is less than the loss rate threshold are designated as standard rapeseed fields.

[0068] Obtain the parameters of the rapeseed harvesting equipment in all the standard rapeseed fields;

[0069] Based on all the parameters of the harvesting equipment, the parameter range of the rapeseed harvesting equipment is calculated to obtain multiple different types of initial parameter ranges;

[0070] Multiple random parameters are generated based on all the initial parameter ranges, and all the random parameters are encoded to obtain multiple initial parameter populations;

[0071] Based on the genetic algorithm, the parameters of all the initial parameter populations are optimized until the maximum number of parameter optimizations is reached, and multiple target device parameters are output.

[0072] Information is exchanged for all the target device parameters, and the target parameter range is calculated based on all the target device parameters that have completed the information exchange.

[0073] The target parameter range is repeatedly updated and iterated until the preset maximum number of iterations is reached, thereby obtaining the optimal parameter range of the rapeseed harvesting equipment.

[0074] Secondly, this application provides an intelligent control device for mechanized rapeseed harvesting based on online monitoring, characterized in that it includes:

[0075] The memory is configured to store instructions; and

[0076] A processor is configured to retrieve the instructions from the memory and, when executing the instructions, to implement a method for intelligent control of mechanized rapeseed harvesting based on online monitoring, as described in any one of the first aspects.

[0077] The above technical solution predicts the maturity period of the target rapeseed field by using rapeseed growth correlation information and a pre-constructed rapeseed growth model. Based on the predicted maturity period, the predicted harvest time is determined. This model-based method identifies a suitable time for rapeseed harvesting, avoiding increased losses due to underripe or overripe rapeseed. Trial harvesting at the predicted time further confirms the optimal harvest time. Furthermore, the adjustment of the harvesting equipment during the trial harvest determines the optimal parameter adjustment range for the subsequent formal harvest, ensuring that the formal harvest occurs at the optimal time and that the harvesting equipment maintains optimal parameters during the harvesting process. In summary, this application provides a method to further reduce rapeseed harvesting losses, minimizing losses and maximizing rapeseed yield.

[0078] Other features and advantages of the embodiments of this application will be described in detail in the following detailed description section. Attached Figure Description

[0079] Figure 1 A flowchart illustrating an intelligent control method for mechanized rapeseed harvesting based on online monitoring, provided in an embodiment of this application;

[0080] Figure 2 A schematic diagram of the rapeseed maturity grade distribution structure of the target rapeseed field provided in the embodiments of this application. Detailed Implementation

[0081] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for illustration and explanation of the embodiments of this application and are not intended to limit the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0082] It should be noted that if the embodiments of this application involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of the components in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0083] Furthermore, if the embodiments of this application involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.

[0084] Figure 1 This illustration schematically shows a flowchart of a method for intelligent control of mechanized rapeseed harvesting based on online monitoring, according to an embodiment of this application. Figure 1 As shown in the figure, this application provides a method for intelligent control of mechanized rapeseed harvesting based on online monitoring. The method may include the following steps:

[0085] S101. Obtain rapeseed growth association information of the target rapeseed field. The rapeseed growth association information includes soil information of the target rapeseed field, meteorological information of the location of the target rapeseed field, and information on the sowing time and rapeseed variety of the rapeseed planted in the target rapeseed field.

[0086] In this embodiment, soil information includes soil pH value, organic matter content, and other information of the target rapeseed field. Meteorological information refers to the rainfall, temperature, and other information of the target rapeseed field location during the rapeseed planting period. Sowing time information refers to the time when rapeseed is sown in the target rapeseed field, such as sowing on March 25th. Rapeseed variety information refers to the type of rapeseed sown in the target rapeseed field, such as Zhongyouza 19, Fengyou 737, and Qinyou 10. Rapeseed growth correlation information is an important factor affecting the rapeseed maturity time. For example, winter rapeseed sown in winter takes 160-290 days from sowing to maturity, while spring rapeseed sown in the Spring Festival takes 85-130 days from sowing to maturity. The growth cycle of Zhongyouza 19 is about 230 days, and the growth cycle of Qinyou 10 is between 230 and 248 days.

[0087] S102. Input rapeseed variety information, meteorological information and soil information into the pre-constructed rapeseed growth model to predict the rapeseed growth cycle, and determine the predicted maturity time period of the target rapeseed field based on the rapeseed growth cycle prediction results.

[0088] In this embodiment, the predicted maturity time period refers to the time it takes for the rapeseed in the target rapeseed field to mature from sowing, which is also the rapeseed growth cycle. For example, the predicted maturity time period is 201 days. A rapeseed growth model is constructed based on a neural network model. Taking a multilayer perceptron (MLP) as an example, historical rapeseed information of the region where the target rapeseed field is located is collected. The historical rapeseed information includes the growth cycles of different varieties of rapeseed under different weather and soil conditions. After data cleaning and labeling, the historical rapeseed information is randomly divided into a training set and a validation set for the rapeseed growth model. The rapeseed growth model includes an input layer for receiving information, a hidden layer for performing nonlinear transformation and feature extraction on the input information, and an output layer for generating the final prediction result. The initial number of nodes is set for the input layer, hidden layer, and output layer according to the feature dimensions of the historical rapeseed information. An appropriate loss function and optimizer are selected. For example, mean squared error (MSE) can be selected as the loss function, and SGD (stochastic gradient descent) can be selected as the optimizer. The loss function is used to measure the difference between the model's prediction result and the true label, and to measure the model's performance. The core function of the optimizer is to update the model parameters based on the gradient of the loss function. This can be used to select an appropriate learning rate or dynamically adjust the learning rate during training, thereby accelerating convergence and improving model training speed. After iteratively training the rapeseed growth model using the training set, the model parameters are continuously optimized, adjusting parameters such as the learning rate and batch size, until the maximum number of iterations is reached. Then, the performance of the trained rapeseed growth model is evaluated using a validation set, with parameters including accuracy and recall. When each parameter reaches a preset threshold, the rapeseed growth model training is complete. Rapeseed variety information, meteorological information, and soil information are input into the trained rapeseed growth model, which then outputs the predicted maturity time for the target rapeseed field.

[0089] S103. Determine the predicted harvest time node of the target rapeseed field based on the sowing time information and the predicted maturity period.

[0090] In this embodiment, the sowing time information is used as the initial node. Adding the predicted maturity period to the initial node yields the predicted harvest time for the target rapeseed field. For example, if the rapeseed planting date is March 21, 2021, and the predicted maturity period is 100 days, then the predicted harvest time is June 29, 2021. The predicted harvest time is the rapeseed maturity time predicted by the rapeseed growth model, and it is also the most suitable time to harvest the rapeseed. Harvesting at the exact maturity time of the rapeseed in the target field can effectively reduce the rapeseed harvesting loss rate. This is because harvesting when the rapeseed is overripe or underripe will lead to a higher harvesting loss rate; therefore, it is necessary to choose an appropriate time for harvesting.

[0091] S104. At the predicted harvest time, remote sensing information of rapeseed in the target rapeseed field is collected by remote sensing drone.

[0092] In this embodiment, at the predicted harvest time, a remote sensing drone equipped with remote sensing devices is controlled to collect remote sensing information of the target rapeseed field. Commonly used remote sensing devices include hyperspectral imagers, thermal infrared sensors, optical cameras, and infrared cameras. At the predicted harvest time, the remote sensing drone is controlled to conduct omnidirectional and regional imaging of the target rapeseed field according to a pre-set monitoring method and path. After the imaging is completed, the acquired remote sensing images are stitched together according to the actual scene of the target rapeseed field based on the real-time positioning information of the remote sensing drone to obtain complete remote sensing information of the target rapeseed field. Rapeseed remote sensing information refers to optical remote sensing images, which are collected by devices such as multispectral cameras and hyperspectral imagers installed inside the remote sensing drone.

[0093] S105. Obtain the rapeseed field size information of the target rapeseed field based on rapeseed remote sensing information, and divide the target rapeseed field into grids according to the rapeseed field size information to obtain multiple grid rapeseed fields.

[0094] In this embodiment, the rapeseed remote sensing information is first subjected to geometric correction, image denoising, and radiometric calibration. Geometric correction refers to the process of eliminating or correcting geometric errors in the remote sensing image, which are image distortions caused by factors such as objective lens distortion and atmospheric refraction. Image denoising mainly aims to remove noise from the rapeseed remote sensing information and improve its accuracy. Radiometric calibration is the process of converting the brightness grayscale values ​​of the rapeseed remote sensing information into absolute radiometric values. Through these steps, the image quality of the rapeseed remote sensing information can be improved, image distortion can be reduced, and errors in subsequent analysis of rapeseed maturity can be reduced. After completing the above preprocessing steps, the rapeseed field size information of the target rapeseed field is calculated based on the rapeseed remote sensing information. Since the edge area of ​​the rapeseed field may be connected to non-rapeseed field areas such as water bodies, roads, field ridges, and buildings when the remote sensing drone collects data on the target rapeseed field, non-rapeseed field areas can be removed based on the rapeseed remote sensing information to more accurately calculate the rapeseed field size information.

[0095] The specific method is as follows: Calculate the red-edge normalized index of the target rapeseed field based on the reflectance of multiple different wavelength bands in the rapeseed remote sensing information. Remove areas with a red-edge normalized index greater than zero. The calculation formula is as follows: N NDVI 705 ＝( ρ 750 -ρ 705 ) / (ρ 750 +ρ 705 ), where ρ 750 ρ represents the reflectivity of the band with a center wavelength of 750nm.705 The reflectance is defined as the band with a center wavelength of 705nm. By calculating the red-edge normalization index, non-rapeseed fields such as water bodies can be removed. Then, based on the rapeseed remote sensing information, non-rapeseed fields such as roads, field ridges, and buildings with reflectance in the green band (550nm) not less than that in the red band (680nm) can be removed, finally obtaining the complete target rapeseed field.

[0096] After obtaining the complete target rapeseed field, the scale information of the rapeseed remote sensing information can be directly obtained using Geographic Information System (GIS) software or other remote sensing data processing software. Alternatively, the scale information of the rapeseed remote sensing information can be determined based on the flight altitude of the UAV and the parameters of the remote sensing equipment. After obtaining the scale information, the image size information of the target rapeseed field in the rapeseed remote sensing information can be directly calculated using GIS software. Finally, the actual size information of the target rapeseed field, i.e., the rapeseed field size information, can be calculated based on the scale information and the image size information.

[0097] After obtaining the rapeseed field size information, the target rapeseed field is divided into grids based on this information. This allows for the division of the target rapeseed field into multiple equally sized grids. The purpose of gridding is to more accurately select a representative rapeseed trial harvesting area within the target rapeseed field. Because the target rapeseed field is large, the rapeseed maturity may vary in different areas, potentially leading to errors in the predicted harvest time. Therefore, it is necessary to select an area within the target rapeseed field with a significant difference in rapeseed maturity and a maturity distribution similar to that of the target field as the trial harvesting area. Trial harvesting is then conducted in this area, and the results are used to determine whether the predicted harvest time is the optimal harvest time.

[0098] S106. Based on the remote sensing information of rapeseed, perform feature analysis on all grid rapeseed fields to obtain the remote sensing features of rapeseed, and calculate the rapeseed maturity of all grid rapeseed fields based on the remote sensing features of rapeseed.

[0099] In this embodiment, the remote sensing features of rapeseed include spectral features and color features. The extraction steps for the spectral features of rapeseed include: using remote sensing image processing software such as ArcGIS and ENVI to extract the reflectance values ​​of the red-edge band, blue band, and near-infrared band in the rapeseed remote sensing information, i.e., spectral reflectance values. Then, using the spectral reflectance values, the Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), and Water Index (WBI) of each grid rapeseed field in the target rapeseed field are calculated, and these indices are integrated to obtain the rapeseed spectral index. The rapeseed spectral index can reflect the maturity of rapeseed. For example, mature rapeseed leaves gradually turn yellow, so its NDVI value will decrease. After rapeseed matures, its water content will also decrease, so its WBI value will decrease accordingly. After normalizing the rapeseed spectral index, the features of the normalized rapeseed spectral index are fused using methods such as linear discriminant analysis, random forest, weighted average, and principal component analysis to obtain the rapeseed spectral features.

[0100] The steps for extracting rapeseed color features include: forming a matrix from the R, G, and B values ​​of image pixels in the rapeseed remote sensing information to obtain the RGB matrix; calculating the covariance matrix of the RGB matrix; calculating the eigenvalues ​​and eigenvectors based on the calculated covariance matrix; sorting the eigenvalues ​​from largest to smallest; forming a matrix from the corresponding eigenvectors; performing principal component transformation to obtain spatial color features; and then using a clustering algorithm to cluster the spatial color features to obtain cluster centers, i.e., the most representative colors. The most representative colors are the rapeseed color features. The principle of using rapeseed color features to analyze rapeseed maturity is that as rapeseed matures, its color gradually changes from light green to yellow, and the rapeseed pods also gradually turn yellowish-brown. Therefore, the rapeseed maturity of each grid rapeseed field can be determined based on the color changes.

[0101] After obtaining the spectral and color features of rapeseed, these features are sequentially input into the rapeseed maturity prediction model. This model can be constructed based on a multi-layer feedforward neural network (BP neural network model). A BP neural network is a multi-layer feedforward network trained using the backpropagation algorithm, typically including an input layer, hidden layers, and an output layer. The input layer can have two neurons, one for the rapeseed spectral features and the other for the rapeseed color features. The output layer can also have multiple neurons, such as four, using four binary digits to represent the rapeseed maturity. The model is trained using a pre-built training set, which includes historical remote sensing features of rapeseed at various maturity levels. The trained model then predicts the rapeseed maturity for each grid rapeseed field.

[0102] S107. Based on the rapeseed maturity, select benchmark grid rapeseed fields from all grid rapeseed fields. With the benchmark grid rapeseed fields as the center and the goal of maximizing the difference in rapeseed maturity, iteratively diffuse and screen the adjacent grid rapeseed fields of the benchmark grid rapeseed fields until the number of selected rapeseed fields is greater than or equal to the preset number threshold, and obtain the rapeseed trial harvesting area.

[0103] In this embodiment, the rapeseed maturity is first classified into maturity levels based on the rapeseed maturity and using a threshold method, resulting in the rapeseed maturity level of all grid rapeseed fields. The rapeseed fields with the highest or lowest rapeseed maturity level are sequentially marked as trial-harvesting rapeseed fields. For any trial-harvesting rapeseed field, the maturity level difference between adjacent rapeseed fields and the trial-harvesting rapeseed field is calculated, with the trial-harvesting rapeseed field as the center. Adjacent rapeseed fields with a maturity level difference greater than or equal to a preset difference threshold are marked as trial-harvesting rapeseed fields. After an adjacent rapeseed field is marked as a trial-harvesting rapeseed field, the marked adjacent rapeseed field is used as the center, and the above trial-harvesting rapeseed field marking steps are repeated. After repeating the trial-harvesting rapeseed field marking steps multiple times, the number of trial-harvesting rapeseed fields is gradually increased until the number of trial-harvesting rapeseed fields is greater than or equal to a preset number threshold, to prevent the purpose of trial harvesting from being unattainable due to too many or too few trial-harvesting rapeseed fields.

[0104] After completing the marking step, adjacent trial-harvested rapeseed fields are integrated to obtain multiple initial trial-harvesting areas. To further ensure that the final selected trial-harvesting areas more accurately represent the target rapeseed fields, a secondary screening of all initial trial-harvesting areas is required. First, a target maturity level matrix for the target rapeseed field is constructed based on the rapeseed maturity levels of different grid rapeseed fields within the target rapeseed field. Similarly, a regional maturity level matrix is ​​constructed for all initial trial-harvesting areas based on the rapeseed maturity levels. Then, a similarity calculation algorithm is used to calculate the similarity between each regional maturity level matrix and the target maturity level matrix, resulting in multiple matrix similarities. The matrix similarity is also a similarity of rapeseed maturity distribution characteristics. The initial trial-harvesting areas with the highest matrix similarity are used as the rapeseed trial-harvesting areas.

[0105] S108. Control the rapeseed harvesting equipment to perform rapeseed trial harvesting tasks in the rapeseed trial harvesting area. During the execution of the rapeseed trial harvesting tasks, use the rapeseed harvesting equipment to count the trial harvesting loss rate in the rapeseed trial harvesting area. Based on the trial harvesting loss rate and using correlation analysis, analyze the total rapeseed loss of the target rapeseed field.

[0106] In this embodiment, when the rapeseed harvesting equipment reaches different grid rapeseed fields within the rapeseed trial harvesting area, an image acquisition device (which can be a high-definition camera) installed outside the rapeseed harvesting equipment is used to acquire rapeseed image information of the grid rapeseed fields. Based on the rapeseed image information and the parameters of the image acquisition device, the rapeseed height information of different grid rapeseed fields is calculated. The equipment parameters of the rapeseed harvesting equipment are set in combination with the rapeseed height information and rapeseed maturity, and the harvesting equipment parameters of different grid rapeseed fields are obtained and stored. While storing the parameters of the harvesting equipment, the system also stores the rapeseed trial harvesting data at the current time point and matches it with the corresponding grid rapeseed field coordinate range. The rapeseed trial harvesting data includes the rapeseed harvesting weight and the amount of rapeseed that has fallen. The rapeseed harvesting weight is obtained through a weight sensor installed inside the rapeseed harvester and refers to the weight of rapeseed harvested in the rapeseed harvester at the current time point. The amount of rapeseed that has fallen is obtained through an image acquisition device installed at the tail of the rapeseed harvester. The image acquisition device (which can be a high-definition camera) acquires images of rapeseed left on the ground after harvesting. The rapeseed density in the rapeseed images is calculated using image recognition technology. Then, the amount of rapeseed that has fallen in each grid rapeseed field is calculated based on the rapeseed density in the rapeseed images. The trial harvesting loss rate of all grid rapeseed fields within the rapeseed trial harvesting area was calculated based on the rapeseed harvest weight and the amount of rapeseed fall. Correlation analysis was performed on the trial harvesting loss rate of each grid rapeseed field and the corresponding rapeseed maturity. The correlation analysis results were obtained by calculating Pearson correlation coefficient, Spearman rank correlation coefficient, Kendall rank correlation coefficient, etc. Then, based on the correlation analysis results, the rapeseed loss rate of all grid rapeseed fields outside the rapeseed trial harvesting area was calculated. Combining the total trial harvesting loss rate and the rapeseed loss rate, the average rapeseed loss rate of the target rapeseed field was calculated. Based on the average rapeseed loss rate and the pre-obtained estimated rapeseed yield, the total rapeseed loss of the target rapeseed field was calculated. The estimated rapeseed yield was calculated based on the historical rapeseed yield of the target rapeseed field. Rapeseed loss rate refers to the loss of rapeseed during harvest due to factors such as unripe or overripe rapeseed, or insufficient performance of rapeseed harvesting equipment. Therefore, rapeseed loss rate is also rapeseed drop rate. By estimating the rapeseed loss rate and adjusting the equipment parameters of rapeseed harvesting equipment accordingly, the rapeseed loss rate can be reduced, rapeseed yield can be increased, and economic benefits for rapeseed growers can be increased.

[0107] S109. Combine the remote sensing characteristics of rapeseed and the total loss of rapeseed to determine the rapeseed maturity node of the target rapeseed field, and judge whether the predicted harvest time node is the optimal harvest node based on the rapeseed maturity node.

[0108] In this embodiment, the predicted harvest time is first determined based on the total rapeseed loss to determine if it is the optimal harvest time. If the predicted harvest time is not the optimal harvest time, the rapeseed maturity time is predicted based on remote sensing features and image information. Specifically, a rapeseed maturity prediction model is constructed based on a neural network model and introduces a self-attention mechanism. The model is iteratively trained using a training set with different root and stem colors, rapeseed height, and remote sensing features, all pre-labeled. After reaching the maximum number of training iterations, the root and stem color features, rapeseed height information, and rapeseed remote sensing features are input into the trained rapeseed maturity prediction model. The model then outputs the predicted rapeseed maturity time for the target rapeseed field. The rapeseed maturity time is added to the current time (predicted harvest time) to obtain the rapeseed maturity time for the target rapeseed field. If the rapeseed maturity time is later than the predicted harvest time, the predicted harvest time is determined to be not the optimal harvest time; if the rapeseed maturity time is not later than the predicted harvest time, the predicted harvest time is determined to be the optimal harvest time. During rapeseed harvesting, the loss rate is closely related to the rapeseed's maturity. Overripe or underripe rapeseed will increase the loss rate. To minimize the loss rate, it is necessary to accurately determine the optimal harvesting time, i.e., the time when the rapeseed is at the appropriate maturity. Harvesting at the optimal harvesting time can effectively reduce the loss rate. At the same time, adjusting the harvesting parameters of the rapeseed harvesting equipment according to the rapeseed's maturity can further reduce the harvesting loss rate. By taking a two-pronged approach of selecting the optimal harvesting time and adjusting the optimal parameters, the rapeseed harvesting loss rate can be minimized and the yield of the target rapeseed field can be increased to the greatest extent.

[0109] S110. If the predicted harvest time node is the optimal harvest node, then combine the rapeseed maturity and trial harvest loss rate and use the parameter optimization algorithm to generate the optimal parameter range of the rapeseed harvesting equipment, and control the rapeseed harvesting equipment to harvest rapeseed in the target rapeseed field according to the optimal parameters.

[0110] In this embodiment, if the predicted harvest time is the optimal harvest time, all grid rapeseed fields in the rapeseed trial harvesting area with a rapeseed loss rate less than the loss rate threshold are first selected to obtain multiple standard rapeseed fields. Based on the coordinate range of each standard rapeseed field, the harvesting equipment parameters within the corresponding standard rapeseed field range are obtained. Multiple random parameters are generated based on all initial parameter intervals, ensuring that all generated random parameters are within their respective initial parameter intervals. Multiple random parameters are generated repeatedly to ensure sufficient diversity in the initial population, preventing the algorithm from prematurely converging to a local optimum. Real-number encoding can be used to construct multiple initial parameter populations. A genetic algorithm is used to optimize the initial parameter populations until the maximum number of parameter optimization iterations is reached, outputting multiple target equipment parameters. Target parameter intervals are constructed based on these multiple target equipment parameters, and the above update and iteration steps are repeated until the preset maximum number of iterations is reached, obtaining the optimal parameter interval for the rapeseed harvesting equipment. The genetic algorithm is a computational model that simulates the biological evolutionary process of natural selection and genetic mechanisms in Darwin's theory of evolution; it is a method for searching for optimal solutions by simulating the natural evolutionary process. This algorithm uses mathematical methods and computer simulations to transform the problem-solving process into a process similar to the crossover and mutation of chromosomes and genes in biological evolution. Subsequently, during the actual harvesting process, the rapeseed harvesting equipment is controlled to make small adjustments to its parameters within the optimal range based on the maturity and height of the rapeseed in the target field. This ensures that the harvesting equipment parameters remain within the optimal range, thereby reducing rapeseed loss.

[0111] In one embodiment, feature analysis is performed on all grid rapeseed fields based on rapeseed remote sensing information to obtain rapeseed remote sensing features. The rapeseed maturity of all grid rapeseed fields is calculated based on the rapeseed remote sensing features, including the following steps:

[0112] Preprocess the remote sensing information of rapeseed in all grid rapeseed fields;

[0113] For any grid rapeseed field, extract the reflectance spectrum from the preprocessed rapeseed remote sensing information;

[0114] The spectral reflectance value is calculated based on the reflectance spectrum, and features are extracted from the spectral reflectance value to obtain the spectral features of rapeseed.

[0115] Principal component transformation was performed on the remote sensing information of rapeseed to obtain the spatial color features of the remote sensing information of rapeseed.

[0116] Clustering algorithms were used to classify spatial color features to obtain the rapeseed color features from remote sensing information.

[0117] The spectral and color characteristics of rapeseed are integrated into remote sensing features of rapeseed in a gridded rapeseed field;

[0118] A rapeseed maturity prediction model was constructed based on a neural network model. All remote sensing features of rapeseed were input into the trained rapeseed maturity prediction model, and the rapeseed maturity of all grid rapeseed fields was output through the rapeseed maturity prediction model.

[0119] In this embodiment, the preprocessing steps for rapeseed remote sensing information include: geometric correction, image denoising, and radiometric calibration. Geometric correction refers to the process of eliminating or correcting geometric errors in the remote sensing image, which are image distortions caused by factors such as objective lens distortion and atmospheric refraction. Image denoising primarily aims to remove noise from the rapeseed remote sensing information, improving its accuracy. Radiometric calibration is the process of converting the brightness grayscale values ​​of the rapeseed remote sensing information into absolute radiometric values. Through these steps, the image quality of the rapeseed remote sensing information can be improved, and image distortion can be reduced.

[0120] The remote sensing features of rapeseed include spectral features and color features. The steps for extracting the spectral features of rapeseed include: using remote sensing image processing software such as ArcGIS and ENVI to extract the reflectance values ​​of the red-edge band, blue band, and near-infrared band in the remote sensing information of rapeseed, i.e., spectral reflectance values. Then, the normalized difference vegetation index (NDVI), enhanced vegetation index (EVI), and water index (WBI) of each grid rapeseed field in the target rapeseed field are calculated using the spectral reflectance values.

[0121] The above index calculation formulas include: NDVI = (NR) / (N+R), where N is the near-infrared reflectance and R is the red light reflectance; EVI = 2.5*((NR) / (N+6R-7.5B+1)), where B represents the blue light reflectance; WBI = (NS) / (N+S), where S is the short-wave infrared reflectance. Rapeseed spectral indices can reflect the maturity of rapeseed. For example, mature rapeseed leaves gradually turn yellow, thus its NDVI value decreases; the water content of mature rapeseed also decreases, thus its WBI value decreases accordingly. After normalizing the rapeseed spectral indices, features of the normalized rapeseed spectral indices are fused using methods such as linear discriminant analysis, random forest, weighted average, and principal component analysis to obtain rapeseed spectral features. Taking principal component analysis as an example, the covariance matrix, eigenvalues, and eigenvectors between indices are calculated, and then principal components are selected for fusion to obtain the rapeseed spectral features.

[0122] The steps for extracting rapeseed color features include: forming a matrix from the R, G, and B values ​​of image pixels in the rapeseed remote sensing information to obtain an RGB matrix; calculating the average value of each variable in the RGB matrix; calculating the covariance of each variable in the RGB matrix based on the calculated average value; constructing the covariance matrix of the RGB matrix based on the calculated covariance; calculating the eigenvalues ​​and eigenvectors of the covariance matrix using the characteristic polynomial; sorting the eigenvalues ​​from largest to smallest; forming a matrix from the corresponding eigenvectors to obtain the PCA transformation matrix, i.e., the spatial color features; and then clustering the spatial color features using a clustering algorithm to obtain cluster centers, i.e., the most representative colors, which are the rapeseed color features. Commonly used clustering algorithms include hierarchical clustering, spectral clustering, and K-means algorithms. Taking the K-means algorithm as an example, k initial cluster centers are selected from the spatial color features. Each pixel is assigned to the nearest cluster center. The center point of each cluster is recalculated, and the assignment and update steps are repeated until the cluster centers no longer change or the predetermined number of iterations is reached. The iteration ends, and the rapeseed color features are obtained. The principle of analyzing rapeseed maturity using rapeseed color characteristics is that as rapeseed matures, its color gradually changes from light green to yellow, and its flower pods also gradually turn yellowish-brown. Therefore, the maturity of rapeseed in each grid rapeseed field can be determined based on the color changes of the rapeseed.

[0123] After obtaining the spectral and color features of rapeseed, these features are sequentially input into the rapeseed maturity prediction model. This model can be constructed based on a multi-layer feedforward neural network (BP neural network model). A BP neural network is a multi-layer feedforward network trained using the backpropagation algorithm, typically including an input layer, hidden layers, and an output layer. The input layer can have two neurons, one for the rapeseed spectral features and the other for the rapeseed color features. The output layer can also have multiple neurons, such as four, using four binary digits to represent the rapeseed maturity. The model is trained using a pre-built training set, which includes historical remote sensing features of rapeseed at various maturity levels. The trained model then predicts the rapeseed maturity for each grid rapeseed field.

[0124] In one embodiment, a rapeseed maturity prediction model is constructed based on a neural network model, and all remote sensing features of rapeseed are input into the trained rapeseed maturity prediction model. The rapeseed maturity prediction model outputs the rapeseed maturity of all grid rapeseed fields, including the following steps:

[0125] A rapeseed maturity prediction model was constructed based on a BP neural network model.

[0126] The rapeseed maturity prediction model was trained using a pre-built rapeseed training set;

[0127] For any grid rapeseed field, the rapeseed spectral characteristics and rapeseed color characteristics of the grid rapeseed field are sequentially input into the rapeseed maturity prediction model to predict the maturity, and the first maturity prediction value and the second maturity prediction value are obtained.

[0128] Kalman filtering is used to filter and correct the first maturity prediction value and the second maturity prediction value;

[0129] The first and second maturity prediction values ​​after filtering correction are fused to obtain the rapeseed maturity of all grid rapeseed fields.

[0130] In this embodiment, a rapeseed maturity prediction model is constructed based on a BP neural network model. A BP neural network is a multi-layer feedforward neural network characterized by forward propagation of signals and backward propagation of errors. It typically includes an input layer, hidden layers, and an output layer. The input layer receives rapeseed spectral and color features and can use two neurons to receive these features respectively. The hidden layer is the core of the neural network, responsible for feature extraction and nonlinear transformation. Through connections and activation functions between multiple neurons, the input data is mapped to the output layer. Softmax (normalized exponential function) can be chosen as the activation function for the hidden layer, and there can be two hidden layers. The output layer neurons calculate the output result through weighted summation and activation functions. There can be four or five neurons in the output layer, and four or five binary numbers can be used to represent rapeseed maturity. Pre-labeled training samples are divided into a rapeseed training set and a rapeseed validation set. The rapeseed maturity prediction model is trained using the rapeseed training set, and parameters such as the learning rate and batch size are dynamically adjusted until the maximum number of iterations is reached. Then, the performance of the trained rapeseed maturity prediction model is evaluated using a rapeseed validation set. The parameters used for evaluation include accuracy and recall. The training of the rapeseed maturity prediction model is complete when all parameters reach preset thresholds. The spectral and color features of rapeseed from each grid rapeseed field are sequentially input into the rapeseed maturity prediction model for maturity prediction. The model predicts a first maturity value and a second maturity value based on the spectral and color features, respectively.

[0131] Then, Kalman filtering is used to correct the first and second maturity prediction values. Specifically, a second prediction is made based on the first and second maturity prediction values, resulting in two secondary maturity prediction values. These secondary maturity prediction values ​​are then used to correct the first and second maturity prediction values. The Kalman gain, a weighting coefficient between 0 and 1, is then calculated using the Kalman filter equation. The corrected first and second maturity prediction values ​​are then fused using the Kalman gain to obtain a more accurate rapeseed maturity level. The Kalman filter algorithm is an optimal recursive filtering method for discrete systems. It can reduce data errors based on data prediction and updates. Using Kalman filtering improves the accuracy of maturity prediction values, resulting in a more accurate rapeseed maturity level. This allows for more reasonable adjustments to the harvesting parameters of the rapeseed harvesting equipment based on the rapeseed maturity level, thereby reducing rapeseed loss.

[0132] In one embodiment, selecting several target grid rapeseed fields from all grid rapeseed fields as rapeseed trial harvesting areas based on rapeseed maturity includes the following steps:

[0133] Based on the rapeseed maturity and using the threshold method, the maturity level of all grid rapeseed fields was divided into different grades, and the rapeseed maturity level of all grid rapeseed fields was obtained.

[0134] If the rapeseed maturity level is the maximum rapeseed maturity level or the rapeseed maturity level is the minimum rapeseed maturity level, then all grid rapeseed fields corresponding to the maximum rapeseed maturity level and the minimum rapeseed maturity level are marked as trial rapeseed fields respectively.

[0135] For any trial rapeseed field, the trial rapeseed field marking step is repeated based on the rapeseed maturity level and centered on the trial rapeseed field, until the number of trial rapeseed fields is greater than or equal to a preset threshold. The trial rapeseed field marking step is as follows:

[0136] Calculate the maturity grade difference between adjacent rapeseed fields and the test rapeseed field, and record adjacent rapeseed fields whose maturity grade difference is greater than or equal to the preset difference threshold as test rapeseed fields;

[0137] By integrating all adjacent rapeseed fields that were previously used for trial harvesting, multiple initial trial harvesting areas were obtained.

[0138] A target maturity level matrix for the target rapeseed field is constructed based on the rapeseed maturity levels of all grid rapeseed fields.

[0139] For any initial trial harvesting area, a regional maturity level matrix for the initial trial harvesting area is constructed based on the rapeseed maturity level;

[0140] The similarity between the target maturity level matrix and the maturity level matrix of all regions is calculated based on the similarity calculation algorithm, resulting in multiple matrix similarities.

[0141] The initial trial cutting area corresponding to the highest data similarity is taken as the rapeseed trial cutting area.

[0142] In this embodiment, refer to Figure 2 First, based on the rapeseed maturity, all grid rapeseed fields are classified into maturity levels using a threshold method. For example, rapeseed maturity of 1-3 is classified as Level 1 maturity, 4-6 as Level 2 maturity, 7-9 as Level 3 maturity, 10-13 as Level 4 maturity, 14-16 as Level 5 maturity, and so on, to obtain the rapeseed maturity level of all grid rapeseed fields. The grid rapeseed fields with the highest or lowest rapeseed maturity level are then marked as trial-harvesting rapeseed fields. For example, if the highest rapeseed maturity level is Level 5 and the lowest is Level 2, then all grid rapeseed fields corresponding to Level 2 and Level 5 maturity are marked as trial-harvesting rapeseed fields.

[0143] In addition, in some cases, if the distance between a certain rapeseed field with the highest or lowest rapeseed maturity level and another rapeseed field with the highest or lowest rapeseed maturity level is less than a preset distance threshold, that is, the distance between the two earliest marked trial rapeseed fields is less than the preset distance threshold, then the straight-line distance between the center point of the two fields and the edge of the target rapeseed field is calculated. The trial rapeseed field with the smaller straight-line distance is marked, and the mark of the rapeseed field with the larger distance is removed. If the distance between the two fields and the edge of the target rapeseed field is equal, then the mark of one trial rapeseed field is randomly removed. This ensures that the distance between the earliest marked trial rapeseed fields is greater than the distance threshold, preventing large-scale overlap between the two initial trial rapeseed areas and thus saving computing power.

[0144] For any given test rapeseed field, taking the test rapeseed field as the center, calculate the maturity level difference between adjacent rapeseed fields and the test rapeseed field. Adjacent rapeseed fields with a maturity level difference greater than or equal to a preset threshold are marked as test rapeseed fields. Adjacent rapeseed fields refer to grid rapeseed fields that overlap with the test rapeseed field in terms of edges. After an adjacent rapeseed field is marked as a test rapeseed field, the above marking process is repeated, using the marked adjacent rapeseed field as the center. After repeating the marking process multiple times, the number of test rapeseed fields gradually increases until it exceeds or equals a preset quantity threshold. The quantity threshold is set to prevent the test rapeseed field from being too large or too small, thus failing to achieve the purpose of the test. The quantity threshold is determined based on the rapeseed field size information of the target rapeseed field and the total number of grid rapeseed fields. In addition, the reason for stopping the rapeseed field marking step when the number is greater than or equal to the quantity threshold is that in some cases, the number of marked trial rapeseed fields may not be exactly equal to the quantity threshold. For example, when the number of marked trial rapeseed fields is one less than the quantity threshold, and the above trial rapeseed field marking step is repeated, there will be two trial rapeseed fields that meet the marking conditions. Therefore, the final number of marked trial rapeseed fields is greater than the quantity threshold.

[0145] After completing the marking step, adjacent trial-harvested rapeseed fields are integrated. Due to the large area of ​​the target rapeseed field and the large number of grid rapeseed fields, multiple initial trial-harvesting areas may be obtained after integrating adjacent trial-harvesting fields. The above steps are to ensure that the difference in rapeseed maturity level between different grid rapeseed fields within the final selected trial-harvesting area is maximized. This is because if the difference in rapeseed maturity level between adjacent grid rapeseed fields within the trial-harvesting area is large, the difference in rapeseed trial-harvesting data between different grid rapeseed fields will also be greater, and the variation of rapeseed harvesting equipment parameters in different grid rapeseed fields will also be greater. This facilitates subsequent analysis of the relationship between rapeseed maturity level and rapeseed trial-harvesting data, as well as the relationship between rapeseed maturity and harvesting equipment parameters, achieving the goal of comprehensively evaluating the target rapeseed field based on the execution results of the rapeseed trial-harvesting task.

[0146] In addition, to ensure that the final selected trial-harvesting areas more accurately represent the target rapeseed fields, a secondary screening of all initial trial-harvesting areas is necessary. First, a target maturity level matrix for the target rapeseed field is constructed based on the rapeseed maturity levels of different grid rapeseed fields within the target rapeseed field. Similarly, a regional maturity level matrix is ​​constructed for all initial trial-harvesting areas based on the rapeseed maturity levels. Then, a similarity calculation algorithm is used to calculate the similarity between each regional maturity level matrix and the target maturity level matrix, resulting in multiple matrix similarities. The initial trial-harvesting area with the highest matrix similarity is selected as the rapeseed trial-harvesting area. Commonly used similarity calculation algorithms include Frobenius norm, cosine similarity, Pearson correlation coefficient, and relative entropy (Kullback-Leibler divergence). Taking the Frobenius norm as an example, first, the regional maturity level matrix is ​​filled with data to ensure that the size of the filled regional maturity level matrix is ​​the same as that of the target maturity level matrix. Then, the difference matrix between the regional maturity level matrix and the target maturity level matrix is ​​calculated. The square root of the sum of squares of each element in the difference matrix is ​​calculated to obtain the Frobenius norm between the regional maturity level matrix and the target maturity level matrix. The Frobenius norm represents the degree of difference between the two matrices. The smaller the Frobenius norm, the more similar the two matrices are. Therefore, the Frobenius norm between the target maturity level matrix and the entire regional maturity level matrix is ​​calculated.

[0147] Next, the grid rapeseed fields located at the edge of the target rapeseed field are designated as edge rapeseed fields. The distances between the center point of each initial trial cutting area and the center points of the edge rapeseed fields in both the horizontal and vertical directions are calculated, and the shortest distance is selected as the edge center distance of that initial trial cutting area. Weights are assigned to the Frobenius norm and the edge center distance, respectively. For example, the weight of the Frobenius norm is 0.8, and the weight of the edge center distance is 0.2. Initial trial cutting areas with both smaller Frobenius norms and smaller edge center distances are selected as trial cutting areas. The purpose of the secondary screening is to select the initial trial harvesting area that is most similar to the rapeseed maturity level distribution of the target rapeseed field and is closest to the edge of the target rapeseed field as the trial harvesting area. The matrix similarity is calculated to make the selected trial harvesting area more representative. The edge center distance is calculated to facilitate the rapeseed harvesting equipment to enter the trial harvesting area. Since the rapeseed harvesting equipment will pre-harvest a part of the rapeseed in the area before entering the trial harvesting area to pave the way for the subsequent entry into the trial harvesting area, if the edge center distance is too large, it will lead to an increase in the area of ​​the pre-harvested area, affecting the subsequent trial harvesting results. This makes the total rapeseed loss of the target rapeseed field calculated based on the trial harvesting results more accurate, and at the same time, it makes the generated optimal parameter interval significantly reduce the total rapeseed loss of the target rapeseed field.

[0148] In one embodiment, controlling rapeseed harvesting equipment to perform rapeseed trial harvesting tasks in a rapeseed trial harvesting area, and statistically analyzing the trial harvesting loss rate of the rapeseed trial harvesting area through the rapeseed harvesting equipment during the execution of the rapeseed trial harvesting task, and analyzing the total rapeseed loss of the target rapeseed field based on the trial harvesting loss rate and using correlation analysis, includes the following steps:

[0149] When the rapeseed harvesting equipment reaches any grid rapeseed field in the rapeseed trial harvesting area, the equipment parameters of the rapeseed harvesting equipment are updated based on the rapeseed maturity of the grid rapeseed field to obtain the harvesting equipment parameters;

[0150] Control the rapeseed harvesting equipment to perform rapeseed trial harvesting tasks in the rapeseed trial harvesting area according to the harvesting equipment parameters, and obtain rapeseed trial harvesting data of all grid rapeseed fields in the rapeseed trial harvesting area;

[0151] The trial harvest loss rate of all grid rapeseed fields in the rapeseed trial harvest area was calculated based on the rapeseed trial harvest data.

[0152] Correlation analysis was performed on the trial harvest loss rate and rapeseed maturity to obtain the correlation analysis results;

[0153] The rapeseed loss rate of all grid rapeseed fields, excluding the rapeseed trial harvesting area, was calculated based on the correlation analysis results.

[0154] The average rapeseed loss rate of the target rapeseed field was calculated by combining the loss rate of all trial harvests and the rapeseed loss rate.

[0155] The total rapeseed loss in the target rapeseed field is calculated based on the average rapeseed loss rate and the pre-obtained estimated rapeseed yield.

[0156] In this embodiment, when the rapeseed harvesting equipment reaches different grid rapeseed fields within the rapeseed trial harvesting area, an image acquisition device (which can be a high-definition camera) installed outside the rapeseed harvesting equipment is used to acquire rapeseed image information of the grid rapeseed fields. Based on the rapeseed image information and the parameters of the image acquisition device, the rapeseed height information of different grid rapeseed fields is calculated. The equipment parameters of the rapeseed harvesting equipment are set in combination with the rapeseed height information and rapeseed maturity to obtain the harvesting equipment parameters for different grid rapeseed fields.

[0157] In addition, the rapeseed harvesting equipment is equipped with a positioning system that stores the coordinate ranges of different grid rapeseed fields within the rapeseed trial harvesting area, as well as the corresponding coordinate ranges. When the positioning system indicates that the harvesting equipment has entered the coordinate range of another grid rapeseed field, the harvesting equipment will immediately determine the rapeseed maturity of the corresponding grid rapeseed field based on the current coordinates. For example, if the grid rapeseed field coordinate range is (60-80, 170-200), and the harvesting equipment's positioning coordinates are (61, 171), it is determined that the harvesting equipment has entered that grid rapeseed field. Then, based on the coordinate range of that grid rapeseed field, it will query the rapeseed maturity corresponding to that grid rapeseed field and begin using the image acquisition device to collect rapeseed image information of that grid rapeseed field. Furthermore, when the image acquisition device calculates suitable harvesting equipment parameters and adjusts them, it will match and store the calculated harvesting equipment parameters with the coordinate range of the grid rapeseed field. Later, the corresponding harvesting equipment parameters can be queried based on the coordinate range.

[0158] While storing the parameters of the harvesting equipment, the system also stores the rapeseed trial harvesting data for the current time point and matches it with the corresponding grid rapeseed field coordinate range. The rapeseed trial harvesting data includes the rapeseed harvesting weight and the amount of rapeseed fallen. The rapeseed harvesting weight is obtained through a weight sensor installed inside the rapeseed harvester, referring to the weight of rapeseed harvested inside the harvester at the current time point. The amount of rapeseed fallen is obtained through an image acquisition device installed at the rear of the rapeseed harvester. The image acquisition device (which can be a high-definition camera) captures images of the rapeseed left on the ground after harvesting, and the rapeseed density in the rapeseed images is calculated using image recognition technology. Specifically, the rapeseed image is enhanced, and then the enhanced rapeseed image is converted to grayscale to obtain a grayscale rapeseed image. Then, using the measurement tools in image processing software, commonly used image processing software includes GIMP (General Image Processing Program) and PIE (Remote Sensing Image Processing Software), the area occupied by rapeseed in the identified grayscale rapeseed image is calculated. Then, the rapeseed density in the rapeseed image is calculated based on the area occupied by rapeseed. Finally, based on the rapeseed density in the rapeseed image, the image acquisition equipment parameters, and the actual area of ​​rapeseed in the grid, the amount of rapeseed that has fallen in each grid rapeseed field is calculated.

[0159] The trial harvesting loss rate for all grid rapeseed fields within the trial harvesting area was calculated based on the rapeseed harvest weight and the amount of rapeseed that fell. Specifically, the rapeseed harvest weight for each trial harvesting area was calculated based on the difference between rapeseed harvest weights stored at adjacent time points. The total rapeseed yield was obtained by adding the harvest weight and the corresponding amount of rapeseed that fell, and then dividing the amount of rapeseed that fell by the total amount of rapeseed. Then, a correlation analysis was performed between the trial harvesting loss rate of each grid rapeseed field and the corresponding rapeseed maturity. The correlation analysis results were obtained by calculating Pearson correlation coefficient, Spearman's rank correlation coefficient, Kendall's rank correlation coefficient, etc. Taking Pearson correlation coefficient as an example, the Pearson correlation coefficient between the trial harvesting loss rate of all grid rapeseed fields and the corresponding rapeseed maturity was calculated using the Pearson correlation coefficient formula, thus obtaining the linear relationship between the trial harvesting loss rate and the corresponding rapeseed maturity, i.e., the correlation analysis results. The Pearson correlation coefficient formula is as follows:

[0160]

[0161] Among them, X i It is a grid-like rapeseed field i The amount of rapeseed harvested. Y is the average rapeseed harvest amount across all grid rapeseed fields. i It is a grid-like rapeseed field i The amount of rapeseed that fell. is the average amount of rapeseed that has fallen across all grid rapeseed fields, and n is the number of grid rapeseed fields.

[0162] Then, based on the correlation analysis results, the rapeseed loss rate of all grid rapeseed fields outside the rapeseed trial harvesting area was calculated. Combining the overall trial harvesting loss rate and the total rapeseed loss rate, the average rapeseed loss rate of the target rapeseed field was calculated. Based on the average rapeseed loss rate and the pre-obtained estimated rapeseed yield, the total rapeseed loss of the target rapeseed field was calculated. The estimated rapeseed yield was calculated based on the historical rapeseed yield of the target rapeseed field. Historical rapeseed yield refers to the yield of the same variety of rapeseed planted in the same month. The average yield of all historical rapeseed yields globally was calculated and used as the estimated rapeseed yield. The average rapeseed loss rate was the average loss rate of each grid rapeseed field in the target rapeseed field. This average loss rate was used as the overall loss rate of the target rapeseed field. Multiplying the average rapeseed loss rate by the estimated rapeseed yield yielded the total rapeseed loss of the target rapeseed field.

[0163] Rapeseed loss rate refers to the loss of rapeseed during harvest due to factors such as unripe or overripe rapeseed, or insufficient performance of rapeseed harvesting equipment. Therefore, rapeseed loss rate is also rapeseed drop rate. By estimating the rapeseed loss rate and adjusting the equipment parameters of rapeseed harvesting equipment accordingly, the rapeseed loss rate can be reduced, rapeseed yield can be increased, and economic benefits for rapeseed growers can be increased.

[0164] In one embodiment, updating the equipment parameters of the rapeseed harvesting equipment based on the rapeseed maturity of the grid rapeseed field to obtain the harvesting equipment parameters includes the following steps:

[0165] Image information of rapeseed in grid rapeseed fields is obtained using image acquisition equipment pre-installed outside the rapeseed harvesting equipment;

[0166] Preprocess rapeseed image information;

[0167] The preprocessed rapeseed image information was binarized using the threshold segmentation method to obtain the target rapeseed image.

[0168] Obtain the equipment parameters of the image acquisition device, and calculate the rapeseed height information of the rapeseed test area based on the equipment parameters and the target rapeseed image;

[0169] By combining rapeseed height information and rapeseed maturity information corresponding to the grid rapeseed fields, the equipment parameters of the rapeseed harvesting equipment are updated to obtain the harvesting equipment parameters.

[0170] In this embodiment, when the rapeseed harvesting equipment reaches different grid rapeseed fields within the rapeseed trial harvesting area, an image acquisition device (which can be a high-definition camera) installed outside the harvesting equipment collects rapeseed image information from the grid rapeseed fields. Based on the rapeseed image information and the parameters of the image acquisition device, the rapeseed height information of different grid rapeseed fields is calculated. The equipment parameters of the rapeseed harvesting equipment are then set in conjunction with the rapeseed height information and rapeseed maturity to obtain the harvesting equipment parameters for different grid rapeseed fields. Specifically, the header height of the rapeseed harvesting equipment is adjusted according to the rapeseed height information. If the rapeseed is tall, the header height needs to be increased accordingly. The cutting speed and feeding speed of the rapeseed harvesting equipment are adjusted according to the rapeseed maturity. If the rapeseed is too mature, the cutting speed and feeding speed should be appropriately reduced to avoid rapeseed detachment due to impact, thus increasing the rapeseed loss rate.

[0171] The analysis of rapeseed image height information includes: removing noise from the rapeseed image using techniques such as mean filtering, median filtering, and Gaussian filtering to improve image quality; converting the rapeseed image to grayscale using threshold segmentation; and segmenting the rapeseed plants from the background to obtain a target rapeseed image without background. Then, connected component analysis is performed on the target rapeseed image. A connected component is a set of interconnected white pixels in the image. Through analysis, each connected component in the image can be identified, and the largest connected component is found, which usually corresponds to the main part of the rapeseed plant. For the largest connected component, its minimum bounding moment is calculated. The minimum bounding moment is the rectangle that completely contains the connected component and has the smallest area. The height of this rectangle (i.e., the vertical side length of the rectangle) can approximate the image height of the rapeseed plant. Then, the equipment parameters of the image acquisition device are obtained, including focal length, principal point coordinates, and distortion coefficients (such as radial and tangential distortion). Based on the equipment parameters, the image height of the rapeseed plant is converted into its actual height to obtain the rapeseed height information.

[0172] In one embodiment, determining the rapeseed maturity node of the target rapeseed field by combining rapeseed remote sensing features and total rapeseed loss, and judging whether the predicted harvest time node is the optimal harvest node based on the rapeseed maturity node includes the following steps:

[0173] If the total loss of rapeseed is less than or equal to the preset loss threshold, then the predicted harvest time node is determined to be the optimal harvest node.

[0174] If the total rapeseed loss exceeds the loss threshold, the rapeseed maturity node in the target rapeseed field is predicted by combining rapeseed image information and rapeseed remote sensing features.

[0175] If the rapeseed matures later than the predicted harvest time, then the predicted harvest time is not the optimal harvest time.

[0176] If the rapeseed matures no later than the predicted harvest time, then the predicted harvest time is determined to be the optimal harvest time.

[0177] In this embodiment, if the predicted total rapeseed loss in the target rapeseed field is less than or equal to a preset loss threshold, the predicted harvest time node can be directly determined as the optimal harvest node because the total rapeseed loss is sufficiently small, and subsequent formal harvesting work can begin. If the total rapeseed loss exceeds the loss threshold, the rapeseed maturity node in the target rapeseed field is re-predicted by combining rapeseed image information and rapeseed remote sensing features. Specifically, the rapeseed root and stem region is first extracted from the rapeseed image information to obtain a rapeseed root and stem image. As rapeseed matures, its roots and stems gradually thicken and harden, and their color may change from light green to dark green or yellow. Therefore, the maturity period of rapeseed can be predicted based on the color characteristics of its roots and stems. Additionally, rapeseed height can also reflect maturity to some extent; for example, rapeseed is generally ready for harvest when its height is between 30-50 cm. A rapeseed maturity prediction model is constructed based on a neural network model and incorporates a self-attention mechanism. The model is iteratively trained using pre-labeled training sets with different root and stem colors, rapeseed heights, and remote sensing features. After reaching the maximum number of training iterations, the root and stem color features, rapeseed height information, and rapeseed remote sensing features are input into the trained rapeseed maturity prediction model. The model then outputs the predicted rapeseed maturity period for the target rapeseed field. The predicted rapeseed maturity period refers to how much time remains until maturity. Adding the current time node (the predicted harvest time node) to the predicted rapeseed maturity period yields the rapeseed maturity node for the target rapeseed field.

[0178] If the rapeseed maturity point is later than the predicted harvest time, it indicates that the optimal harvest time has not yet been reached. Therefore, the predicted harvest time is not considered the optimal harvest time. Once the optimal harvest time is reached, the rapeseed maturity prediction model is used to predict the overall maturity of the target rapeseed field, obtaining the target maturity. Then, all grid rapeseed fields in the rapeseed trial harvesting area with a rapeseed loss rate less than the loss rate threshold are selected as target grid rapeseed fields. The harvesting equipment parameters corresponding to all target grid rapeseed fields are then obtained, and the average value of all harvesting equipment parameters is calculated. Based on the average equipment parameters and the target maturity, the average equipment parameters are adjusted to obtain the optimal average equipment parameters. The rapeseed harvesting equipment is then controlled according to the optimal average equipment parameters for formal rapeseed harvesting. If the rapeseed maturity point is not later than the predicted harvest time, it indicates that the current time may be the optimal harvest time, or it may have already exceeded the optimal harvest time. Therefore, the predicted harvest time is directly determined as the optimal harvest time.

[0179] In one embodiment, predicting the rapeseed maturity node of a target rapeseed field by combining rapeseed image information and rapeseed remote sensing features includes the following steps:

[0180] A rapeseed maturity prediction model was constructed based on a neural network model and a self-attention mechanism, and the model was trained using a pre-built training set.

[0181] The root and stem regions of rapeseed images are identified using a target detection algorithm. Based on the identification results, the rapeseed images are segmented to obtain rapeseed root and stem images.

[0182] Clustering algorithms were used to extract the root and stem color features from rapeseed root and stem images;

[0183] The root and stem color features, rapeseed height information, and rapeseed remote sensing features are input into the trained rapeseed maturity prediction model, and the predicted rapeseed maturity period of the target rapeseed field is output through the rapeseed maturity prediction model.

[0184] The predicted harvest time is adjusted based on the predicted rapeseed maturity date to obtain the rapeseed maturity date of the target rapeseed field.

[0185] In this embodiment, a rapeseed maturity prediction model is constructed based on a neural network model and a self-attention mechanism. Taking a multilayer perceptron (MLP) model as an example, an input layer is designed to receive root and stem color features, rapeseed height information, and rapeseed remote sensing features. A self-attention layer is introduced into the model so that the model can focus on the relationships and importance between different features. Weights are assigned to the root and stem color features, rapeseed height information, and rapeseed remote sensing features according to the relationships and importance between different features. The self-attention layer is constructed based on the self-attention mechanism. By performing a linear transformation on the input features, a query, key, and value are generated. The dot product between the query vector and the key vector is calculated to obtain the attention score matrix. The dot product result is divided by a scaling factor (which can be the square root of the key vector) to stabilize the gradient. The scaled attention score matrix is ​​then transformed into probability values ​​between [0,1] using a normalization function to obtain the attention weights. The normalization function can be the softmax function. The obtained attention weights are multiplied by the value vector to complete the weight allocation for the root and stem color features, rapeseed height information, and rapeseed remote sensing features. The hidden layer of the rapeseed maturity prediction model uses an activation function to map the input data to the output layer. Softmax (normalized exponential function) can be chosen as the activation function for the hidden layer. Two hidden layers can be configured. A pre-labeled sample set, with varying root and stem colors, rapeseed height, and remote sensing features, is divided into a training set and a validation set. The rapeseed maturity prediction model is iteratively trained until the maximum number of training iterations is reached. Cross-validation can be used to evaluate the performance of the rapeseed maturity prediction model. Cross-validation includes K-fold cross-validation and leave-one-out cross-validation. Taking K-fold cross-validation as an example, the sample set is divided into K subsets. Each time, K-1 subsets are selected as the training set, and the remaining subset is used as the validation set. This process is repeated K times, each time selecting a different subset as the validation set. Leave-one-out cross-validation, on the other hand, leaves only one sample as the validation set each time, and the rest as the training set. This process is repeated until every sample has been used as the validation set. In each cross-validation iteration, the model is trained using the training set and its performance is evaluated using the validation set. Then, adjust the model parameters or select the optimal model based on the performance metrics (such as accuracy, recall, F1 score, etc.) on the validation set. After all cross-validation iterations are completed, calculate the average performance metric of the model on the validation set. If the average performance metric is greater than or equal to the preset metric threshold, it indicates that the rapeseed maturity prediction model has been trained successfully.

[0186] Object detection algorithms are used to identify the root and stem regions in rapeseed images. Commonly used algorithms include the R-CNN series and the YOLO series. Taking YOLO (a deep learning-based object detection algorithm) as an example, the rapeseed image information is input into a neural network. Through operations such as convolution and pooling, the keyframes are divided into multiple grids. Each grid predicts whether a root and stem region exists. Based on the root and stem region identification results, the root and stem region is segmented from the rapeseed image information, resulting in the rapeseed root and stem image. First, the color space of the rapeseed root and stem image is converted from RGB (red, green, blue) to HSV (hue, saturation, brightness). The RGB values ​​are normalized to a range, and then the hue, saturation, and brightness are calculated based on R, G, and B, thus completing the color space conversion. After color space conversion, clustering algorithms are used to identify the dominant color in the rapeseed root and stem image. A commonly used clustering algorithm is k-means clustering, which clusters pixels in the rapeseed root and stem image, grouping pixels of similar colors into the same class, calculating the center of each class (i.e., the dominant color), extracting the center color of each class, and calculating the frequency of each color. The color with the highest frequency is taken as the dominant color, i.e., the root and stem color feature. The root and stem color feature, rapeseed height information, and rapeseed remote sensing features are input into a trained rapeseed maturity prediction model. The model outputs the predicted rapeseed maturity date for the target rapeseed field. The predicted rapeseed maturity date refers to how much time is left until the rapeseed matures. For example, if the rapeseed maturity date is 10 days, the current time node (predicted harvest time node) is added to the rapeseed maturity date to obtain the rapeseed maturity node for the target rapeseed field.

[0187] In one embodiment, generating the optimal parameter range for the rapeseed harvesting equipment by combining rapeseed maturity and trial harvesting loss rate and using a parameter optimization algorithm includes the following steps:

[0188] All grid rapeseed fields in the rapeseed trial harvest area with a rapeseed loss rate less than the loss rate threshold were designated as standard rapeseed fields.

[0189] Obtain the harvesting equipment parameters for all standard rapeseed fields;

[0190] Based on all harvesting equipment parameters, the parameter range of the rapeseed harvesting equipment is calculated, resulting in multiple different types of initial parameter ranges.

[0191] Multiple random parameters are generated based on all initial parameter ranges, and all random parameters are encoded to obtain multiple initial parameter populations;

[0192] Based on the genetic algorithm, the parameters of all initial parameter populations are optimized until the maximum number of parameter optimizations is reached, and multiple target device parameters are output.

[0193] Exchange information on all target equipment parameters, and calculate the target parameter range based on all target equipment parameters that have been exchanged.

[0194] The target parameter range is repeatedly updated and iterated until the preset maximum number of iterations is reached, thus obtaining the optimal parameter range for the rapeseed harvesting equipment.

[0195] In this embodiment, all grid rapeseed fields with a rapeseed loss rate less than the loss rate threshold in the rapeseed trial harvesting area are first screened out to obtain multiple standard rapeseed fields. The harvesting equipment parameters of the rapeseed harvesting equipment within the corresponding standard rapeseed field range are obtained according to the coordinate range of each standard rapeseed field. The harvesting equipment parameters include header height, cutter speed, and feeding speed. The maximum and minimum values ​​of each type of parameter are screened from all harvesting equipment parameters, namely, maximum header height, minimum header height, fastest cutter speed, slowest cutter speed, fastest feeding speed, and slowest feeding speed. Based on the maximum and minimum values ​​of the above parameters, multiple different types of initial parameter ranges are constructed. For example, the initial parameter range for header height is 8-14 cm, the initial parameter range for cutter speed is 1.8-2.7 m / s, and the initial parameter range for feeding speed is 0.6-1.3 m / s. First, multiple random parameters are generated based on the entire initial parameter range. These random parameters all lie within their respective initial parameter ranges. Multiple random parameters are generated repeatedly to ensure sufficient diversity in the initial population, preventing premature convergence to local optima. Real-number encoding can be used, where each gene is a real number representing a specific parameter value. Each random parameter is then treated as a gene for an individual, and each individual consists of multiple genes. Several individuals form a population, resulting in multiple initial parameter populations. The fitness of each individual in the initial parameter population is calculated using a pre-constructed fitness function. Based on the individual fitness, a selection operator is used to select the next generation population. A tournament selection operator can be chosen. The core idea of ​​tournament selection is to randomly select several chromosomes (individuals) and choose the individual with the highest fitness from this set of chromosomes for the next generation. This process is repeated until a preset population size is reached. Then, a crossover operator is used to perform a crossover operation on the next generation population to obtain offspring populations. Crossover operators can include single-point crossover, multi-point crossover, and uniform crossover. Taking uniform crossover as an example, each gene position is exchanged with a certain probability to generate new offspring individuals. After the crossover operation is completed, the offspring population is mutated using the mutation operator to obtain the updated parameter population. Differential mutation or equal mutation can be used. Taking differential mutation as an example, an individual is randomly selected as the base individual, and two different individuals are randomly selected. The difference vector between them is calculated, the difference vector is multiplied by a scaling factor, and then added to the base individual to generate a new individual.The above optimization iterations are continued on the updated parameter population until the maximum number of iterations is reached. Multiple optimal equipment parameters of different types are output, i.e., multiple target equipment parameters. Information exchange is conducted among all target equipment parameters to achieve co-evolution. This can be achieved through mutation operations, i.e., introducing new gene information to facilitate information exchange among the target equipment parameters. Similarly, the maximum and minimum values ​​of different types of parameters in the target equipment parameters are selected to construct target parameter intervals. The above iterative steps are repeated on the target parameter intervals to further narrow them down until the preset maximum number of iterations is reached, obtaining the optimal parameter interval for the rapeseed harvesting equipment. Subsequently, during the formal harvesting process, the rapeseed harvesting equipment is controlled to make small-scale parameter adjustments within the optimal parameter interval based on the rapeseed maturity and height in the target rapeseed field. This ensures that the harvesting equipment parameters of the rapeseed harvesting equipment can be continuously maintained within the optimal parameter interval, thereby reducing the rapeseed loss rate.

[0196] This application also discloses an intelligent control device for mechanized rapeseed harvesting based on online monitoring, characterized in that it includes:

[0197] The memory is configured to store instructions; and

[0198] The processor is configured to retrieve instructions from memory and, when executing the instructions, to implement a method for intelligent control of mechanized rapeseed harvesting based on online monitoring, according to any of the above.

[0199] The processor can be a central processing unit (CPU). Of course, depending on the actual use, it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), off-the-shelf programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc., and this application does not limit it in this regard.

[0200] The memory can be an internal storage unit of a computer device, such as a hard disk or RAM, or an external storage device, such as a plug-in hard disk, smart memory card (SMC), secure digital card (SD), or flash memory card (FC) provided on the computer device. Furthermore, the memory can be a combination of internal storage units and external storage devices of a computer device. The memory is used to store computer programs and other programs and data required by the computer device. The memory can also be used to temporarily store data that has been output or will be output. This application does not limit this.

[0201] This application also provides a machine-readable storage medium storing instructions that cause a machine to execute the above-described method for intelligent control of mechanized rapeseed harvesting based on online monitoring.

[0202] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0203] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0204] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0205] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0206] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0207] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0208] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0209] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0210] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for intelligent control of mechanized rapeseed harvesting based on online monitoring, characterized in that, The method includes the following steps: Obtain rapeseed growth association information of the target rapeseed field, including soil information of the target rapeseed field, meteorological information of the location of the target rapeseed field, and sowing time and rapeseed variety information of the rapeseed planted in the target rapeseed field; The rapeseed variety information, meteorological information, and soil information are input into a pre-constructed rapeseed growth model to predict the rapeseed growth cycle. Based on the rapeseed growth cycle prediction results, the predicted maturity time period of the target rapeseed field is determined. The predicted harvest time node of the target rapeseed field is determined based on the sowing time information and the predicted maturity time period; At the predicted harvest time, remote sensing information of rapeseed in the target rapeseed field is collected by a remote sensing drone; Based on the remote sensing information of rapeseed, the size information of the target rapeseed field is obtained, and the target rapeseed field is divided into grids according to the size information of the rapeseed field to obtain multiple grid rapeseed fields; Based on the rapeseed remote sensing information, feature analysis is performed on all the grid rapeseed fields to obtain rapeseed remote sensing features, and the rapeseed maturity of all the grid rapeseed fields is calculated based on the rapeseed remote sensing features. Based on the rapeseed maturity, a benchmark grid rapeseed field is selected from all the grid rapeseed fields. Taking the benchmark grid rapeseed field as the center and the maximization of the rapeseed maturity difference as the selection objective, the adjacent grid rapeseed fields of the benchmark grid rapeseed field are iteratively diffused and screened until the number of selected rapeseed fields is greater than or equal to a preset number threshold, and the rapeseed trial harvesting area is obtained. The rapeseed harvesting equipment is controlled to perform rapeseed trial harvesting tasks in the rapeseed trial harvesting area. During the execution of the rapeseed trial harvesting tasks, the rapeseed harvesting equipment is used to count the trial harvesting loss rate in the rapeseed trial harvesting area. Based on the trial harvesting loss rate and using correlation analysis, the total rapeseed loss of the target rapeseed field is analyzed. The rapeseed maturity node of the target rapeseed field is determined by combining the rapeseed remote sensing features and the total rapeseed loss, and the predicted harvest time node is determined as the optimal harvest time node based on the rapeseed maturity node. If the predicted harvest time node is the optimal harvest node, then the optimal parameter range of the rapeseed harvesting equipment is generated by combining the rapeseed maturity and the trial harvesting loss rate and using a parameter optimization algorithm, and the rapeseed harvesting equipment is controlled to harvest rapeseed in the target rapeseed field according to the optimal parameters.

2. The method according to claim 1, characterized in that, The step of performing feature analysis on all the grid rapeseed fields based on the rapeseed remote sensing information to obtain rapeseed remote sensing features, and calculating the rapeseed maturity of all the grid rapeseed fields based on the rapeseed remote sensing features, includes the following steps: Preprocess the remote sensing information of all the rapeseed fields in the grid; For any of the grid rapeseed fields, extract the reflectance spectrum from the preprocessed rapeseed remote sensing information; The spectral reflectance value is calculated based on the reflected spectrum, and the spectral reflectance value is used to extract features to obtain the spectral features of rapeseed. Principal component transformation is performed on the rapeseed remote sensing information to obtain the spatial color features of the rapeseed remote sensing information; Clustering algorithms are used to classify the spatial color features to obtain the rapeseed color features of the rapeseed remote sensing information; The rapeseed spectral features and rapeseed color features are integrated into the rapeseed remote sensing features of the grid rapeseed field; A rapeseed maturity prediction model is constructed based on a neural network model, and all the remote sensing features of rapeseed are input into the trained rapeseed maturity prediction model. The rapeseed maturity prediction model outputs the rapeseed maturity of all the grid rapeseed fields.

3. The method according to claim 2, characterized in that, The process of constructing a rapeseed maturity prediction model based on a neural network model, inputting all the remote sensing features of rapeseed into the trained rapeseed maturity prediction model, and outputting the rapeseed maturity of all the grid rapeseed fields through the rapeseed maturity prediction model includes the following steps: A rapeseed maturity prediction model was constructed based on a BP neural network model. The rapeseed maturity prediction model is trained using a pre-constructed rapeseed training set; For any of the grid rapeseed fields, the rapeseed spectral characteristics and rapeseed color characteristics of the grid rapeseed fields are sequentially input into the rapeseed maturity prediction model to predict maturity, and a first maturity prediction value and a second maturity prediction value are obtained. The first maturity prediction value and the second maturity prediction value are filtered and corrected using Kalman filtering; The first maturity prediction value and the second maturity prediction value after filtering and correction are fused to obtain the rapeseed maturity of all the grid rapeseed fields.

4. The method according to claim 1, characterized in that, The step of selecting several target grid rapeseed fields as rapeseed trial harvesting areas from all the grid rapeseed fields based on the rapeseed maturity includes the following steps: Based on the rapeseed maturity and using the threshold method, the maturity levels of all the grid rapeseed fields are divided into different grades to obtain the rapeseed maturity grades of all the grid rapeseed fields. If the rapeseed maturity level is the maximum rapeseed maturity level or the rapeseed maturity level is the minimum rapeseed maturity level, then all the grid rapeseed fields corresponding to the maximum rapeseed maturity level and the minimum rapeseed maturity level are marked as trial rapeseed fields respectively. For any of the aforementioned trial-harvested rapeseed fields, the trial-harvested rapeseed field marking step is repeatedly performed based on the rapeseed maturity level and centered on the trial-harvested rapeseed field, until the number of trial-harvested rapeseed fields is greater than or equal to a preset threshold. The trial-harvested rapeseed field marking step is as follows: Calculate the maturity grade difference between the adjacent rapeseed fields and the test rapeseed field, and record the adjacent rapeseed fields whose maturity grade difference is greater than or equal to a preset difference threshold as test rapeseed fields; By integrating all the adjacent rapeseed fields that were previously tested, multiple initial test areas were obtained; Construct a target maturity level matrix for the target rapeseed field based on the rapeseed maturity levels of all the grid rapeseed fields; For any of the initial trial harvesting areas, a regional maturity level matrix for the initial trial harvesting area is constructed based on the rapeseed maturity level; The similarity between the target maturity level matrix and all the regional maturity level matrices is calculated based on the similarity calculation algorithm to obtain multiple matrix similarities; The initial trial cutting area corresponding to the largest matrix similarity is taken as the rapeseed trial cutting area.

5. The method according to claim 1, characterized in that, The controlled rapeseed harvesting equipment performs a rapeseed trial harvesting task in the rapeseed trial harvesting area. During the execution of the rapeseed trial harvesting task, the rapeseed harvesting equipment statistically analyzes the trial harvesting loss rate of the rapeseed trial harvesting area. Based on the trial harvesting loss rate and using correlation analysis, the total rapeseed loss of the target rapeseed field is analyzed, including the following steps: When the rapeseed harvesting equipment reaches any of the grid rapeseed fields in the rapeseed trial harvesting area, the equipment parameters of the rapeseed harvesting equipment are updated based on the rapeseed maturity of the grid rapeseed field to obtain the harvesting equipment parameters; The rapeseed harvesting equipment is controlled to perform rapeseed trial harvesting tasks in the rapeseed trial harvesting area according to the harvesting equipment parameters, and rapeseed trial harvesting data of all the grid rapeseed fields in the rapeseed trial harvesting area are obtained; The trial harvesting loss rate of all grid rapeseed fields in the rapeseed trial harvesting area was calculated based on the rapeseed trial harvesting data. A correlation analysis was performed on the trial harvest loss rate and the rapeseed maturity to obtain the correlation analysis results. The rapeseed loss rate of all grid rapeseed fields, excluding the rapeseed trial harvesting area, was calculated based on the correlation analysis results. The average rapeseed loss rate of the target rapeseed field is calculated by combining all the trial harvest loss rates and the rapeseed loss rate. The total rapeseed loss of the target rapeseed field is calculated based on the average rapeseed loss rate and the pre-obtained estimated rapeseed yield.

6. The method according to claim 5, characterized in that, The process of updating the equipment parameters of the rapeseed harvesting equipment based on the rapeseed maturity in the grid rapeseed field to obtain the harvesting equipment parameters includes the following steps: Image information of rapeseed in the grid rapeseed field is obtained using an image acquisition device pre-installed outside the rapeseed harvesting equipment; Preprocess the rapeseed image information; The preprocessed rapeseed image information is binarized using a threshold segmentation method to obtain the target rapeseed image. Obtain the device parameters of the image acquisition device, and calculate the rapeseed height information of the rapeseed test cutting area based on the device parameters and the target rapeseed image; The rapeseed harvesting equipment parameters are updated by combining the rapeseed height information and the rapeseed maturity corresponding to the grid rapeseed field, thus obtaining the harvesting equipment parameters.

7. The method according to claim 6, characterized in that, The process of determining the rapeseed maturity node of the target rapeseed field by combining the rapeseed remote sensing features and the total rapeseed loss, and determining whether the predicted harvest time node is the optimal harvest time node based on the rapeseed maturity node, includes the following steps: If the total loss of rapeseed is less than or equal to a preset loss threshold, then the predicted harvest time node is determined to be the optimal harvest node. If the total loss of rapeseed is greater than the loss threshold, then the rapeseed maturity node of the target rapeseed field is predicted by combining the rapeseed image information and the rapeseed remote sensing features. If the rapeseed maturity point is later than the predicted harvest time point, then the predicted harvest time point is determined not to be the optimal harvest time point; If the rapeseed maturity point is not later than the predicted harvest time point, then the predicted harvest time point is determined to be the optimal harvest time point.

8. The method according to claim 7, characterized in that, The step of predicting the rapeseed maturity node in the target rapeseed field by combining the rapeseed image information and the rapeseed remote sensing features includes the following steps: A rapeseed maturity prediction model was constructed based on a neural network model and a self-attention mechanism, and the model was trained using a pre-constructed training set. The root and stem regions of the rapeseed image information are identified using a target detection algorithm. Based on the root and stem region identification results, the rapeseed image information is segmented to obtain the rapeseed root and stem image in the rapeseed image information. Clustering algorithms were used to extract the root and stem color features of the rapeseed root and stem images; The root and stem color features, rapeseed height information, and rapeseed remote sensing features are input into the trained rapeseed maturity prediction model, and the predicted rapeseed maturity period of the target rapeseed field is output through the rapeseed maturity prediction model. The predicted harvest time is corrected based on the predicted rapeseed maturity date to obtain the rapeseed maturity date of the target rapeseed field.

9. The method according to claim 6, characterized in that, The step of combining the rapeseed maturity and the trial harvesting loss rate and using a parameter optimization algorithm to generate the optimal parameter range for the rapeseed harvesting equipment includes the following steps: All grid rapeseed fields in the rapeseed trial harvesting area whose rapeseed loss rate is less than the loss rate threshold are designated as standard rapeseed fields. Obtain the parameters of the rapeseed harvesting equipment in all the standard rapeseed fields; Based on all the parameters of the harvesting equipment, the parameter range of the rapeseed harvesting equipment is calculated to obtain multiple different types of initial parameter ranges; Multiple random parameters are generated based on all the initial parameter ranges, and all the random parameters are encoded to obtain multiple initial parameter populations; Based on the genetic algorithm, the parameters of all the initial parameter populations are optimized until the maximum number of parameter optimizations is reached, and multiple target device parameters are output. Information is exchanged for all the target device parameters, and the target parameter range is calculated based on all the target device parameters that have completed the information exchange. The target parameter range is repeatedly updated and iterated until the preset maximum number of iterations is reached, thereby obtaining the optimal parameter range of the rapeseed harvesting equipment.

10. An intelligent control device for mechanized rapeseed harvesting based on online monitoring, characterized in that, include: The memory is configured to store instructions; as well as The processor is configured to retrieve the instructions from the memory and, when executing the instructions, to implement the method for intelligent control of mechanized rapeseed harvesting based on online monitoring according to any one of claims 1 to 9.

Images