Automatic control system and method for rotary tillage depth of rotary tiller based on soil hardness detection

Abstract

The present invention discloses a system and method for automatically controlling the tillage depth of a rotary tiller based on soil hardness detection, and relates to the field of automatic control systems. The present invention optimizes and adjusts the tillage depth of the rotary tiller by establishing a mapping model between the soil hardness factor of cultivated land and the working energy consumption and tool wear of the rotary tiller, so that the tool wear level difference and energy consumption level data of the rotary tiller after the work is completed can meet the requirements, thereby improving the working effect of the rotary tiller; wherein, the tool wear level difference and energy consumption level data of the rotary tillers corresponding to multiple cultivated lands before and after work are calculated, which provides a quantitative judgment basis for subsequent judgment on whether the tillage depth data of the rotary tiller needs to be adjusted, and also provides an adjustment direction for adjusting the tillage depth data of the rotary tiller; constructing a first final SVM classification model and a second final SVM classification model, thereby improving the feasibility of implementation and ease of operation.

Description

Technical Field

[0001] The present invention belongs to the field of automatic control systems, and in particular, relates to a system and method for automatically controlling the tillage depth of a rotary tiller based on soil hardness detection. Background Art

[0002] The current rotary tiller equipment cannot adjust its tillage depth according to the soil conditions under different soil conditions. Therefore, when faced with moist soil, hard soil, or soil with a large amount of gravel or lumps, due to improper pre-setting of the rotary tillage depth, the rotary tiller blades may be damaged, energy consumption may increase, or even the rotary tiller may be damaged, which may lead to interruption of the tillage process or poor results, and increased maintenance costs of the rotary tiller. Summary of the Invention

[0003] In response to the problems in the related art, the present invention proposes a system and method for automatically controlling the tillage depth of a rotary tiller based on soil hardness detection to overcome the above-mentioned technical problems existing in the existing related art.

[0004] To solve the above technical problems, the present invention is achieved through the following technical solutions:

[0005] The present invention is a method for automatically controlling the tillage depth of a rotary tiller based on soil hardness detection, comprising the following steps:

[0006] S1. Collect average soil moisture data, average soil hardness data, and average soil texture grade data for multiple cultivated fields to obtain an average soil moisture dataset, an average soil hardness dataset, and an average soil texture grade dataset; and collect tillage depth requirement data for crops corresponding to the multiple cultivated fields to obtain a crop tillage depth dataset;

[0007] S2. Calculating the tool wear level difference and energy consumption level data of rotary tillers corresponding to a plurality of cultivated fields before and after operation to obtain an initial tool wear level difference data set and an initial energy consumption level data set;

[0008] S3, using the initial tool wear level difference data set, the initial energy consumption level data set, the soil moisture average data set, the soil hardness average data set, and the soil texture level average data set to construct a first final SVM classification model and a second final SVM classification model; and optimizing and adjusting the initial tillage depth data of the rotary tillers corresponding to the plurality of cultivated fields in combination with the initial tool wear level difference data set, the initial energy consumption level data set, the soil moisture average data set, the soil hardness average data set, the soil texture level average data set, the first final SVM classification model, the second final SVM classification model, and the crop tillage depth data set to obtain a final tillage depth data set;

[0009] S4. adjusting parameters of rotary tillers corresponding to the plurality of cultivated fields according to the final tillage depth data set;

[0010] Since the soil moisture, soil hardness and soil texture of the cultivated land have a great influence on the rotary tillage of the cultivated land by the rotary tiller, this scheme collects the average soil moisture data, average soil hardness data and average soil texture grade data of multiple cultivated lands, which provides data support for the subsequent construction of the first final SVM classification model and the second final SVM classification model that can map the soil moisture, soil hardness, soil texture and tillage depth data to the tool wear grade difference data and energy consumption grade data; then, by collecting the tillage depth requirement data of the crops corresponding to multiple cultivated lands, the adjustment range is determined for the subsequent adjustment of the tillage depth data of the rotary tiller, so that the adjusted tillage depth data can better meet the actual needs; the tool wear degree and energy consumption data of the rotary tiller can reflect the quality of the current rotary tiller's tillage depth data. The smaller the tool wear degree and energy consumption data of the rotary tiller are, the better the working effect of the current rotary tiller is; based on the above, this scheme calculates the tool wear degree and energy consumption data of the rotary tiller corresponding to multiple cultivated lands. The wear level difference and energy consumption level data before and after work provide a quantitative basis for determining whether the tillage depth data of the rotary tiller needs to be adjusted, and also provide an adjustment direction for adjusting the tillage depth data of the rotary tiller; by constructing the first final SVM classification model and the second final SVM classification model, a mapping model between soil data and rotary tillage depth data to tool wear level difference and energy consumption level data is established, so that when the tillage depth data of the rotary tiller is subsequently adjusted, there is no need to actually measure the tool wear level difference and energy consumption level data of the rotary tiller to detect the quality of the adjusted data, thereby improving the feasibility and ease of operation of this solution; by adjusting the tillage depth data of the rotary tiller whose tool wear level difference and energy consumption level data do not meet the requirements, the tool wear level difference and energy consumption level data of the rotary tiller after the work is completed can meet the requirements, thereby improving the working effect of the rotary tiller and reducing the maintenance cost of the rotary tiller.

[0011] Preferably, the S1 comprises the following steps:

[0012] S11, set a plurality of arable lands of the same size that need to be rotary tilled and corresponding rotary tillers of the same type and quality, and obtain the arable land sample set a={a1, a2, ..., a i ,...,a a′} and rotary tiller sample set, a i represents the set i-th piece of cultivated land sample that needs to be rotary tilled, and a′ represents the total number of set cultivated land samples that need to be rotary tilled;

[0013] S12, measuring the cultivated land sample set a={a1, a2, ..., a i ,...,a a′ The average soil moisture data, average soil hardness data and average soil texture grade data of each cultivated land sample in} are obtained to obtain the average soil moisture data set Soil hardness average dataset and soil texture class average dataset Respectively represent the average soil moisture data, average soil hardness data, and average soil texture grade data corresponding to the i-th block of cultivated land sample that needs to be rotary tilled in the cultivated land sample set;

[0014] S13, set the cultivated land sample set a={a1, a2, ..., a i ,...,a a′ The tillage depth requirement data of the crops planted in each piece of cultivated land sample in} is obtained to obtain the crop tillage depth dataset Represents the cultivated land sample set a={a1,a2,...,a i ,...,a a′} the tillage depth requirement data of the crops planted on the i-th cultivated land sample;

[0015] By setting up multiple plots of arable land of the same size that need to be rotary tilled and corresponding rotary tillers of the same type and quality, it is ensured that subsequent data measurements and adjustments to the tillage depth data are carried out under the same premise, making the final adjustment results more reliable and accurate.

[0016] Preferably, the S12 includes the following steps:

[0017] S121, in the cultivated land sample set a={a1, a2, ..., a i ,...,a a′}, multiple soil collection points are evenly set on each cultivated land sample to obtain the soil collection point matrix b; as follows,

[0018]

[0019] Among them, b ij represents the j-th soil collection point set on the i-th cultivated land sample in the cultivated land sample set;

[0020] S122, collecting soil samples from the corresponding cultivated land samples in the cultivated land sample set according to the soil collection point matrix b, to obtain a soil sample matrix as follows,

[0021]

[0022] in, represents the soil sample collected at the j-th soil collection point set on the i-th cultivated land sample in the cultivated land sample set;

[0023] S123, using a resistive soil moisture sensor to measure the soil sample matrix The humidity data of each soil sample is measured to obtain a soil humidity data matrix; a soil hardness meter is used to measure the soil sample matrix The hardness data of each soil sample is measured to obtain a soil hardness data matrix; the average value of each row of data in the soil moisture data matrix and the soil hardness data matrix is calculated to obtain a soil moisture average data set and soil hardness average dataset

[0024] Measuring the soil sample matrix The texture grade average data of each soil sample in each cultivated land sample is obtained to obtain the soil texture grade average data set;

[0025] By evenly setting multiple soil collection points on each cultivated land sample to collect soil samples, the sporadic nature of soil data obtained in subsequent measurements is reduced, thereby improving the accuracy of the measurement.

[0026] Preferably, the soil sample matrix is measured in S123 The texture grade average data of each soil sample in each cultivated land sample is obtained, and the soil texture grade average data set is obtained, which includes the following steps:

[0027] S1231, measuring the soil sample matrix The mass percentage data of sand, silt and clay in each soil sample are obtained to obtain the mass percentage data set matrix as follows,

[0028]

[0029] in, represent the mass percentage data of sand particles, silt particles and clay particles in the soil sample collected at the j-th soil collection point set on the i-th cultivated land sample in the cultivated land sample set respectively;

[0030] Calculate the mass percentage dataset matrix The average of the mass percentage data of sand, silt and clay in each row of data is obtained to obtain the mass percentage average data set

[0031] c={(c 11 ,c 12 ,c 13 ),(c 21 ,c 22 ,c23 ),...,(c i1 ,c i2 ,c i3 ),...,(c a′1 ,c a′2 ,c a′3 )},c i1 、c i2 、c i3 They respectively represent the average value of the mass percentage data of sand particles, silt particles and clay particles in the soil samples collected at each soil collection point on the i-th cultivated land sample in the cultivated land sample set; the calculation formula is as follows:

[0032]

[0033] S1232, set the soil texture grade set; perform clustering operation on the mass percentage average data set to obtain the mass percentage average data classification matrix set c i ′ represents the i-th mass percentage average data classification matrix obtained by clustering the mass percentage average data set; as follows,

[0034]

[0035] Among them, c i ' j1 、c i ' j2 、c i ' j3 Respectively represent c i The mass percentage data of sand, silt and clay in the jth soil sample in d i Indicates clustering to c i The total number of soil samples in ′;

[0036] S1233, classifying the mass percentage average data matrix set according to the soil texture grade set The soil texture grade of each mass percentage average data in the classification matrix is set to obtain the soil texture grade average data set

[0037]

[0038] By clustering the mass percentage average data set, similar soil textures are clustered into one category, and then the soil texture data clustered into one category are set to the same level, thereby reducing the complexity of the soil texture level data and facilitating subsequent analysis.

[0039] Preferably, in S1232, a clustering operation is performed on the mass percentage average data set to obtain a mass percentage average data classification matrix set. The following steps are involved:

[0040] S12321. Establish the first squirrel population d1′ i represents the i-th squirrel in the first squirrel population, Represents the size of the first squirrel population; Set the maximum number of iterations of the first squirrel population to The current number of iterations is They are respectively recorded as the first maximum number of iterations and the first current number of iterations; the search space dimension of the first squirrel population is

[0041] S12322, multiple times from the mass percentage data set matrix Randomly select several mass percentage data sets as the initial position matrix of each squirrel in the first squirrel population, and get the first initial position matrix set e 1i represents the initial position matrix of the i-th squirrel in the first squirrel population; as follows,

[0042]

[0043] Among them, e 1ij1 、e 1ij2 、e 1ij3 represent the components of the mass percentage data dimensions of sand, silt, and clay in the jth soil sample in the initial position matrix of the i-th squirrel in the first squirrel population;

[0044] S12323, calculate the Euclidean distance between each mass percentage average data in the mass percentage average data set and each row of data in the initial position matrix of the i-th squirrel in the first squirrel population, and obtain the Euclidean distance matrix e i ';as follows,

[0045]

[0046] Among them, e i ' i′j It represents the Euclidean distance between the i′th data in the mass percentage average data set and the jth row data in the initial position matrix of the i-th squirrel in the first squirrel population; the calculation formula is as follows:

[0047]

[0048] Where c i′1 、c i′2 、c i′3represent the average values of the mass percentages of sand, silt and clay in the soil samples collected at each soil collection point on the i′th cultivated land sample in the cultivated land sample set;

[0049] According to the Euclidean distance matrix e i 'Construct the fitness function of the i-th squirrel in the first squirrel population as follows,

[0050]

[0051] S12324, start iteration, before iteration, set the first current iteration number Set to 1; the fitness function of the i-th squirrel is used in the first iteration Calculate the first initial position matrix set The fitness value of each initial position matrix in the first fitness value set is obtained to obtain a first fitness value set; the maximum fitness value in the first fitness value set and the corresponding initial position matrix are respectively used as the first global optimal fitness and the first global optimal position; the first initial position matrix set is adjusted according to the first global optimal fitness and the first global optimal position. After the update is completed, the first current iteration number Add 1 and enter the next iteration;

[0052] In each other round of iteration, the fitness function of the i-th squirrel is used Calculate the fitness value of each position matrix updated in the previous round of iteration to obtain a second fitness value set; use the maximum fitness value and the corresponding position matrix in the second fitness value set as the second global optimal fitness and the second global optimal position respectively; continue to update each position matrix updated in the previous round of iteration according to the second global optimal fitness and the second global optimal position; after the update is completed, the first current iteration number is Add 1 and enter the next iteration;

[0053] S12325, when the When , the iteration is stopped and the first final global optimal position is output; the mass percentage average data set is classified according to the first final global optimal position to obtain the mass percentage average data classification matrix set

[0054] The squirrel optimization algorithm can effectively explore and utilize food resources in the search space by simulating the foraging strategy and gliding motion of squirrels, thereby achieving strong optimization ability. Based on this, this scheme uses the squirrel optimization algorithm to perform multiple iterative optimizations on the multiple cluster center data of the mass percentage average data set, and uses the overall discreteness of the mass percentage average data set as its fitness function. As the iteration proceeds, the overall discreteness of the mass percentage average data set becomes smaller and smaller, indicating that the multiple cluster center data of the mass percentage average data set found are more accurate.

[0055] Preferably, said S2 comprises the following steps:

[0056] S21, setting a plurality of rotary tiller tool wear levels to obtain a tool wear level set; measuring tool wear level data of each rotary tiller in a rotary tiller sample set before operation based on the tool wear level set to obtain an initial tool wear level data set;

[0057] According to the crop tillage depth dataset Set the initial tillage depth data to obtain the initial tillage depth data set represents the initial tillage depth data of the i-th rotary tiller in the rotary tiller sample set;

[0058] S22. Set multiple energy consumption levels and the tillage depth data of each rotary tiller is fixed during operation to obtain an energy consumption level set; make each rotary tiller in the rotary tiller sample set start working, and after the work is completed, measure the tool wear level data and energy consumption level data of each rotary tiller in the rotary tiller sample set according to the tool wear level set and the energy consumption level set to obtain a final tool wear level data set and an initial energy consumption level data set. represents the energy consumption level data of the i-th rotary tiller in the rotary tiller sample set during operation;

[0059] The final tool wear level data set is subtracted from the initial tool wear level data set to obtain the initial tool wear level difference data set. The initial tool wear level difference data of the rotary tiller corresponding to the i-th piece of cultivated land sample that needs to be rotary tilled is set;

[0060] By setting multiple wear levels and energy consumption levels for rotary tiller tools, the complexity of the subsequently collected wear data and energy consumption data for the rotary tiller tools is reduced; by subtracting the final tool wear level data set from the initial tool wear level data set, the amount of wear on the rotary tiller tools during operation can be reflected.

[0061] Preferably, the step S3 includes the following steps:

[0062] S31, setting a wear level difference data threshold and an energy consumption level data threshold;

[0063] S32, when the initial tool wear level difference data set There is an initial tool wear level difference data greater than or equal to the wear level difference data threshold or initial energy consumption level data set When the energy consumption level data in the data is greater than or equal to the energy consumption level data threshold, the crop tillage depth dataset is satisfied. Under the premise of , the tillage depth data of the corresponding rotary tiller is adjusted until the initial tool wear level difference data set is There is no initial tool wear level difference data greater than or equal to the wear level difference data threshold and the initial energy consumption level data set The final tillage depth dataset is obtained until there is no energy consumption level data greater than or equal to the energy consumption level data threshold. represents the final tillage depth data of the i-th rotary tiller in the rotary tiller sample set;

[0064] Otherwise, there is no need to adjust the tillage depth data of the rotary tiller;

[0065] By setting the wear level difference data threshold and the energy consumption level data threshold, a comparison basis is set for the wear level difference data and the energy consumption level data of the rotary tiller, providing an adjustment direction for subsequent adjustment of the rotary tillage depth data of the rotary tiller.

[0066] Preferably, adjusting the tillage depth data of the corresponding rotary tiller in S32 includes the following steps:

[0067] S321, constructing a first initial SVM classification model and a second initial SVM classification model;

[0068] S322, training and testing a first initial SVM classification model using the initial tillage depth dataset, the initial tool wear level difference dataset, the soil moisture average dataset, the soil hardness average dataset, and the soil texture level average dataset to obtain a first final SVM classification model;

[0069] The second initial SVM classification model is trained and tested using the initial tillage depth dataset, the initial energy consumption level dataset, the soil moisture average dataset, the soil hardness average dataset, and the soil texture level average dataset to obtain a second final SVM classification model;

[0070] S323: Adjust the tillage depth data of the corresponding rotary tiller in conjunction with the first final SVM classification model and the second final SVM classification model to obtain a final tillage depth data set.

[0071] Preferably, in S322, the initial tillage depth dataset, the initial tool wear level difference dataset, the soil moisture average dataset, the soil hardness average dataset, and the soil texture level average dataset are used to train and test the first initial SVM classification model to obtain the first final SVM classification model, including the following steps:

[0072] S32211. Set a first training data ratio; divide the initial plowing depth dataset, the initial tool wear level difference dataset, the soil moisture average dataset, the soil hardness average dataset, and the soil texture level average dataset according to the first training data ratio to obtain a first initial plowing depth training dataset, an initial tool wear level difference training dataset, a first soil moisture average training dataset, a first soil hardness average training dataset, a first soil texture level average training dataset, a first initial plowing depth test dataset, an initial tool wear level difference test dataset, a first soil moisture average test dataset, a first soil hardness average test dataset, and a first soil texture level average test dataset;

[0073] S32212. Set a first training error threshold; input the first initial tillage depth training data set, the initial tool wear level difference training data set, the first soil moisture average training data set, the first soil hardness average training data set, and the first soil texture level average training data set into a first initial SVM classification model for training; during the training process, when the training error is less than the first training error threshold, stop training to obtain a first trained SVM classification model; otherwise, continue training until the training error is less than the first training error threshold;

[0074] S32213. Set a first test accuracy threshold; input the first initial plowing depth test data set, the initial tool wear level difference test data set, the first soil moisture average test data set, the first soil hardness average test data set, and the first soil texture level average test data set into the first trained SVM classification model for testing; after the test is completed, obtain a first test accuracy; when the first test accuracy is greater than or equal to the first test accuracy threshold, use the first trained SVM classification model as the first final SVM classification model; otherwise, return to S3212 to continue training the first trained SVM classification model until the first test accuracy is greater than or equal to the first test accuracy threshold;

[0075] Preferably, in S322, the initial tillage depth dataset, the initial energy consumption level dataset, the soil moisture average dataset, the soil hardness average dataset, and the soil texture level average dataset are used to train and test the second initial SVM classification model to obtain the second final SVM classification model, including the following steps:

[0076] S32221. Set a second training data ratio; divide the initial plowing depth dataset, the initial energy consumption level dataset, the soil moisture average dataset, the soil hardness average dataset, and the soil texture level average dataset according to the second training data ratio to obtain a second initial plowing depth training dataset, an initial energy consumption level training dataset, a second soil moisture average training dataset, a second soil hardness average training dataset, a second soil texture level average training dataset, a second initial plowing depth test dataset, a second initial energy consumption level test dataset, a second soil moisture average test dataset, a second soil hardness average test dataset, and a second soil texture level average test dataset;

[0077] S32222. Set a second training error threshold; input the second initial tillage depth training data set, the initial energy consumption level training data set, the second soil moisture average training data set, the second soil hardness average training data set, and the second soil texture level average training data set into a second initial SVM classification model for training; during the training process, when the training error is less than the second training error threshold, stop training to obtain a second trained SVM classification model; otherwise, continue training until the training error is less than the second training error threshold;

[0078] S32223. Set a second test accuracy threshold; input the second initial plowing depth test data set, the initial energy consumption level test data set, the second soil moisture average test data set, the second soil hardness average test data set, and the second soil texture level average test data set into the second trained SVM classification model for testing; after the test is completed, obtain the second test accuracy; when the second test accuracy is greater than or equal to the second test accuracy threshold, use the second trained SVM classification model as the second final SVM classification model; otherwise, return to S3222 to continue training the second trained SVM classification model until the second test accuracy is greater than or equal to the second test accuracy threshold;

[0079] By using the initial plowing depth dataset, the initial tool wear level difference dataset, the soil moisture average dataset, the soil hardness average dataset and the soil texture level average dataset to train and test the first initial SVM classification model, the obtained first final SVM classification model has a good mapping ability between the initial plowing depth dataset, the soil moisture average dataset, the soil hardness average dataset and the soil texture level average dataset and the initial tool wear level difference dataset; then using the initial plowing depth dataset, the initial energy consumption level dataset, the soil moisture average dataset, the soil hardness average dataset and the soil texture level average dataset to train and test the second initial SVM classification model, the obtained second final SVM classification model has a good mapping ability between the initial plowing depth dataset, the soil moisture average dataset, the soil hardness average dataset and the soil texture level average dataset and the initial energy consumption level dataset, thereby providing an accurate classification model for the subsequent plowing depth data of the rotary tiller.

[0080] Preferably, the S323 includes the following steps:

[0081] S3231: Set the tillage depth data to be adjusted and the corresponding tool wear level difference data f1′ to be optimized and the energy consumption level data f2′ to be optimized. Acquire the average soil moisture data, average soil hardness data, and average soil texture data corresponding to the tillage depth data to be adjusted by combining the average soil moisture data set, the average soil hardness data set, and the average soil texture data set. Record these data as average soil moisture adjustment data, average soil hardness adjustment data, and average soil texture adjustment data, respectively.

[0082] S3232, build the second squirrel population d2′ i represents the i-th squirrel in the second squirrel population, Represents the size of the second squirrel population; Set the maximum number of iterations of the second squirrel population to The current number of iterations is are respectively recorded as the second maximum number of iterations and the second current number of iterations; the search space dimension of the second squirrel population is 1 dimension;

[0083] S3233, setting the initial position of each squirrel in the second squirrel population according to the tillage depth data to be adjusted, and obtaining the initial position set f i represents the initial position of the i-th squirrel in the second squirrel population; the calculation formula is as follows,

[0084] f i =f min +rand i ·(fmax -f min );

[0085] Where, f min 、f max Respectively represent the lower limit and upper limit of the tillage depth data to be adjusted, rand i For f i Generate a random number between 0 and 1;

[0086] S3234: Construct the fitness function of the second squirrel population based on the tool wear level difference data f1′ to be optimized and the energy consumption level data f2′ to be optimized. as follows,

[0087]

[0088] In the formula, β is a positive number, indicating the correction parameter;

[0089] S3235, start iteration, before iteration, set the second current iteration number Set to 1; the fitness function of the second squirrel population is used in the first iteration And cooperate with the first final SVM classification model and the second final SVM classification model to calculate the initial position set The fitness value of each initial position in the third fitness value set is obtained; the maximum fitness value and the corresponding initial position in the third fitness value set are respectively used as the third global optimal fitness and the third global optimal position; the initial position set is adjusted according to the third global optimal fitness and the third global optimal position. After the update is completed, the second current iteration number Add 1 and enter the next iteration;

[0090] The fitness function of the second squirrel population is used in each other iteration process And cooperate with the first final SVM classification model and the second final SVM classification model to calculate the fitness value of each position updated in the previous round of iteration to obtain a fourth fitness value set; the maximum fitness value in the fourth fitness value set and the corresponding position matrix are respectively used as the fourth global optimal fitness and the fourth global optimal position; each position updated in the previous round of iteration is continued to be updated according to the fourth global optimal fitness and the fourth global optimal position; after the update is completed, the second current iteration number is Add 1 and enter the next iteration;

[0091] S3236, when the When , the iteration is stopped and the second final global optimal position is output; the corresponding initial tillage depth data in the initial tillage depth data set is replaced by the second final global optimal position to obtain the final tillage depth data set

[0092] The squirrel optimization algorithm is used to perform multiple iterative optimizations on the tillage depth of the rotary tiller that does not meet the requirements, and the fitness function is constructed using the tool wear level difference data and energy consumption level data of the rotary tiller. Therefore, as the iteration proceeds, the tool wear level difference data and energy consumption level data of the rotary tiller will become smaller and smaller, and eventually meet the set requirements.

[0093] Preferably, said S4 comprises the following steps:

[0094] S41, adjusting the parameters of each rotary tiller in the rotary tiller sample set to change the rotary tillage depth, and after the adjustment is completed, using ultrasound to actually measure the rotary tillage depth of each rotary tiller in the rotary tiller sample set to obtain an actual rotary tillage depth dataset;

[0095] S42, setting a rotary tillage depth error threshold; when the difference between the corresponding rotary tillage depth data in the actual rotary tillage depth data set and the final tillage depth data set is greater than or equal to the rotary tillage depth error threshold, returning to S41 to continue parameter adjustment until the difference between the corresponding rotary tillage depth data in the actual rotary tillage depth data set and the final tillage depth data set is greater than or equal to the rotary tillage depth error threshold; otherwise, there is no need to return to S41 to continue parameter adjustment;

[0096] The parameters of the rotary tiller are adjusted through multiple cycles so that the error between the actual tillage depth data of the rotary tiller and the final tillage depth data obtained by previous optimization meets the requirements, thereby making the working effect of the adjusted rotary tiller better.

[0097] The automatic control system for the tillage depth of a rotary tiller based on soil hardness detection includes a cultivated land soil data measurement and calculation module, a crop tillage depth requirement data collection module, a rotary tiller data measurement and calculation module, a classification model construction module, a tillage depth data adjustment module, and a rotary tiller parameter adjustment module.

[0098] The cultivated land soil data measurement and calculation module is used to collect average soil moisture data, average soil hardness data and average soil texture grade data of multiple cultivated lands to obtain an average soil moisture data set, an average soil hardness data set and an average soil texture grade data set;

[0099] The crop tillage depth requirement data collection module is used to collect tillage depth requirement data of crops corresponding to a plurality of cultivated lands to obtain a crop tillage depth data set;

[0100] The rotary tiller data measurement and calculation module is used to calculate the wear level difference and energy consumption level data of the rotary tillers corresponding to multiple cultivated fields before and after operation, and obtain an initial tool wear level difference data set and an initial energy consumption level data set;

[0101] The classification model construction module is used to construct a first final SVM classification model and a second final SVM classification model using the initial tool wear level difference data set, the initial energy consumption level data set, the soil moisture average data set, the soil hardness average data set, and the soil texture level average data set;

[0102] The tillage depth data adjustment module is used to optimize and adjust the initial tillage depth data of the rotary tiller corresponding to the plurality of cultivated fields in conjunction with the first final SVM classification model, the second final SVM classification model and the crop tillage depth dataset to obtain a final tillage depth dataset;

[0103] The rotary tiller parameter adjustment module is used to adjust the parameters of the rotary tillers corresponding to the plurality of cultivated fields according to the final tillage depth data set.

[0104] The present invention has the following beneficial effects:

[0105] 1. In the present invention, a mapping model between the soil hardness factor of the cultivated land and the working energy consumption and tool wear of the rotary tiller is established to optimize and adjust the rotary tillage depth of the rotary tiller, so that the tool wear level difference and energy consumption level data of the rotary tiller after the work is completed can meet the requirements, thereby improving the working effect of the rotary tiller; wherein, the tool wear level difference and energy consumption level data of the rotary tiller corresponding to multiple cultivated lands before and after work are calculated, which provides a quantitative judgment basis for subsequent determination of whether the tillage depth data of the rotary tiller needs to be adjusted, and also provides an adjustment direction for adjusting the tillage depth data of the rotary tiller; constructing the first final SVM classification model and the second final SVM classification model improves the feasibility of implementation and ease of operation.

[0106] 2. In the present invention, the squirrel optimization algorithm is used to perform multiple iterative optimizations on the multiple cluster center data of the mass percentage average data set and the tillage depth data of the rotary tiller. As the iteration proceeds, the overall discreteness of the mass percentage average data set, the tool wear level difference data, and the energy consumption level data become smaller and smaller.

[0107] 3. In the present invention, by setting the wear level difference data threshold and the energy consumption level data threshold, a comparison basis is set for the wear level difference data and the energy consumption level data of the rotary tiller, which provides an adjustment direction for the subsequent adjustment of the rotary tillage depth data of the rotary tiller.

[0108] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0109] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, they can also obtain drawings based on these drawings without paying any creative work.

[0110] Figure 1 The present invention is a schematic diagram of a process for regulating the tillage depth of a rotary tiller by an automatic control system for the tillage depth of a rotary tiller based on soil hardness detection. DETAILED DESCRIPTION

[0111] The following will clearly and completely describe the technical solutions in the embodiments of the invention in conjunction with the accompanying drawings. Obviously, the embodiments described are only part of the embodiments of the invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0112] In the description of the present invention, it should be understood that the terms "opening", "upper", "lower", "top", "middle", "inside" and the like indicating orientation or positional relationship are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the invention.

[0113] Example 1

[0114] This embodiment is a method for automatically controlling the tillage depth of a rotary tiller based on soil hardness detection, comprising the following steps:

[0115] S1. Collect average soil moisture data, average soil hardness data, and average soil texture grade data for multiple cultivated fields to obtain an average soil moisture dataset, an average soil hardness dataset, and an average soil texture grade dataset; and collect tillage depth requirement data for crops corresponding to the multiple cultivated fields to obtain a crop tillage depth dataset;

[0116] Said S1 comprises the following steps:

[0117] S11, set a plurality of arable lands of the same size that need to be rotary tilled and corresponding rotary tillers of the same type and quality, and obtain the arable land sample set a={a1, a2, ..., a i ,...,a a′} and rotary tiller sample set, a irepresents the set i-th piece of cultivated land sample that needs to be rotary tilled, and a′ represents the total number of set cultivated land samples that need to be rotary tilled;

[0118] S12, measuring the cultivated land sample set a={a1, a2, ..., a i ,...,a a′ The average soil moisture data, average soil hardness data and average soil texture grade data of each cultivated land sample in} are obtained to obtain the average soil moisture data set a~1={a~ 11 ,a~ 12 ,...,a~ 1i ,...,a~ 1a′}, soil hardness average data set a~2={a~ 21 ,a~ 22 ,...,a~ 2i ,...,a~ 2a′} and soil texture class average dataset Respectively represent the average soil moisture data, average soil hardness data, and average soil texture grade data corresponding to the i-th block of cultivated land sample that needs to be rotary tilled in the cultivated land sample set;

[0119] The S12 includes the following steps:

[0120] S121, in the cultivated land sample set a={a1, a2, ..., a i ,...,a a′}, multiple soil collection points are evenly set on each cultivated land sample to obtain the soil collection point matrix b; as follows,

[0121]

[0122] Among them, b ij represents the j-th soil collection point set on the i-th cultivated land sample in the cultivated land sample set;

[0123] S122, collecting soil samples from the corresponding cultivated land samples in the cultivated land sample set according to the soil collection point matrix b, to obtain a soil sample matrix as follows,

[0124]

[0125] in, represents the soil sample collected at the j-th soil collection point set on the i-th cultivated land sample in the cultivated land sample set;

[0126] S123, using a resistive soil moisture sensor to measure the soil sample matrix The humidity data of each soil sample is measured to obtain a soil humidity data matrix; a soil hardness meter is used to measure the soil sample matrix The hardness data of each soil sample is measured to obtain a soil hardness data matrix; the average value of each row of data in the soil moisture data matrix and the soil hardness data matrix is calculated to obtain a soil moisture average data set and soil hardness average dataset

[0127] Measuring the soil sample matrix The texture grade average data of each soil sample in each cultivated land sample is obtained to obtain the soil texture grade average data set;

[0128] S123 measures the soil sample matrix The texture grade average data of each soil sample in each cultivated land sample is obtained, and the soil texture grade average data set is obtained, which includes the following steps:

[0129] S1231, measuring the soil sample matrix The mass percentage data of sand, silt and clay in each soil sample are obtained to obtain the mass percentage data set matrix as follows,

[0130]

[0131] in, represent the mass percentage data of sand particles, silt particles and clay particles in the soil sample collected at the j-th soil collection point set on the i-th cultivated land sample in the cultivated land sample set respectively;

[0132] Calculate the mass percentage dataset matrix The average of the mass percentage data of sand, silt and clay in each row of data is obtained to obtain the mass percentage average data set

[0133] c={(c 11 ,c 12 ,c 13 ),(c 21 ,c 22 ,c 23 ),...,(c i1 ,c i2 ,c i3 ),...,(c a′1 ,c a′2 ,c a′3 )},c i1 、c i2 、c i3They respectively represent the average value of the mass percentage data of sand particles, silt particles and clay particles in the soil samples collected at each soil collection point on the i-th cultivated land sample in the cultivated land sample set; the calculation formula is as follows:

[0134]

[0135] S1232, set the soil texture grade set; perform clustering operation on the mass percentage average data set to obtain the mass percentage average data classification matrix set c i ′ represents the i-th mass percentage average data classification matrix obtained by clustering the mass percentage average data set; as follows,

[0136]

[0137] Among them, c i ' j1 、c i ' j2 、c i ' j3 Respectively represent c i The mass percentage data of sand, silt and clay in the jth soil sample in d i Indicates clustering to c i The total number of soil samples in ′;

[0138] In S1232, a clustering operation is performed on the mass percentage average data set to obtain a mass percentage average data classification matrix set. The following steps are involved:

[0139] S12321. Establish the first squirrel population d1′ i represents the i-th squirrel in the first squirrel population, Represents the size of the first squirrel population; Set the maximum number of iterations of the first squirrel population to The current number of iterations is They are respectively recorded as the first maximum number of iterations and the first current number of iterations; the search space dimension of the first squirrel population is

[0140] S12322, multiple times from the mass percentage data set matrix Randomly select several mass percentage data sets as the initial position matrix of each squirrel in the first squirrel population, and get the first initial position matrix set e 1i represents the initial position matrix of the i-th squirrel in the first squirrel population; as follows,

[0141]

[0142] Among them, e 1ij1 、e 1ij2 、e 1ij3 represent the components of the mass percentage data dimensions of sand, silt, and clay in the jth soil sample in the initial position matrix of the i-th squirrel in the first squirrel population;

[0143] S12323, calculate the Euclidean distance between each mass percentage average data in the mass percentage average data set and each row of data in the initial position matrix of the i-th squirrel in the first squirrel population, and obtain the Euclidean distance matrix e i ';as follows,

[0144]

[0145] Among them, e i ' i′j It represents the Euclidean distance between the i′th data in the mass percentage average data set and the jth row data in the initial position matrix of the i-th squirrel in the first squirrel population; the calculation formula is as follows:

[0146]

[0147] Where c i′1 、c i′2 、c i′3 represent the average values of the mass percentages of sand, silt and clay in the soil samples collected at each soil collection point on the i′th cultivated land sample in the cultivated land sample set;

[0148] According to the Euclidean distance matrix e i 'Construct the fitness function of the i-th squirrel in the first squirrel population as follows,

[0149]

[0150] S12324, start iteration, before iteration, set the first current iteration number Set to 1; the fitness function of the i-th squirrel is used in the first iteration Calculate the first initial position matrix set The fitness value of each initial position matrix in the first fitness value set is obtained to obtain a first fitness value set; the maximum fitness value in the first fitness value set and the corresponding initial position matrix are respectively used as the first global optimal fitness and the first global optimal position; the first initial position matrix set is adjusted according to the first global optimal fitness and the first global optimal position. After the update is completed, the first current iteration number Add 1 and enter the next iteration;

[0151] In each other round of iteration, the fitness function of the i-th squirrel is used Calculate the fitness value of each position matrix updated in the previous round of iteration to obtain a second fitness value set; use the maximum fitness value and the corresponding position matrix in the second fitness value set as the second global optimal fitness and the second global optimal position respectively; continue to update each position matrix updated in the previous round of iteration according to the second global optimal fitness and the second global optimal position; after the update is completed, the first current iteration number is Add 1 and enter the next iteration;

[0152] S12325, when the When , the iteration is stopped and the first final global optimal position is output; the mass percentage average data set is classified according to the first final global optimal position to obtain the mass percentage average data classification matrix set

[0153] S1233, classifying the mass percentage average data matrix set according to the soil texture grade set The soil texture grade of each mass percentage average data in the classification matrix is set to obtain the soil texture grade average data set

[0154]

[0155] S13, set the cultivated land sample set a={a1, a2, ..., a i ,...,a a′ The tillage depth requirement data of the crops planted in each piece of cultivated land sample in} is obtained to obtain the crop tillage depth dataset Represents the cultivated land sample set a={a1,a2,...,a i ,...,a a′} the tillage depth requirement data of the crops planted on the i-th cultivated land sample;

[0156] S2. Calculating the tool wear level difference and energy consumption level data of rotary tillers corresponding to a plurality of cultivated fields before and after operation to obtain an initial tool wear level difference data set and an initial energy consumption level data set;

[0157] The S2 comprises the following steps:

[0158] S21, setting a plurality of rotary tiller tool wear levels to obtain a tool wear level set; measuring tool wear level data of each rotary tiller in a rotary tiller sample set before operation based on the tool wear level set to obtain an initial tool wear level data set;

[0159] According to the crop tillage depth dataset Set the initial tillage depth data to obtain the initial tillage depth data set represents the initial tillage depth data of the i-th rotary tiller in the rotary tiller sample set;

[0160] S22. Set multiple energy consumption levels and the tillage depth data of each rotary tiller is fixed during operation to obtain an energy consumption level set; make each rotary tiller in the rotary tiller sample set start working, and after the work is completed, measure the tool wear level data and energy consumption level data of each rotary tiller in the rotary tiller sample set according to the tool wear level set and the energy consumption level set to obtain a final tool wear level data set and an initial energy consumption level data set. represents the energy consumption level data of the i-th rotary tiller in the rotary tiller sample set during operation;

[0161] The final tool wear level data set is subtracted from the initial tool wear level data set to obtain the initial tool wear level difference data set. The initial tool wear level difference data of the rotary tiller corresponding to the i-th piece of cultivated land sample that needs to be rotary tilled is set;

[0162] S3, using the initial tool wear level difference data set, the initial energy consumption level data set, the soil moisture average data set, the soil hardness average data set, and the soil texture level average data set to construct a first final SVM classification model and a second final SVM classification model; and optimizing and adjusting the initial tillage depth data of the rotary tillers corresponding to the plurality of cultivated fields in combination with the initial tool wear level difference data set, the initial energy consumption level data set, the soil moisture average data set, the soil hardness average data set, the soil texture level average data set, the first final SVM classification model, the second final SVM classification model, and the crop tillage depth data set to obtain a final tillage depth data set;

[0163] The S3 includes the following steps:

[0164] S31, setting a wear level difference data threshold and an energy consumption level data threshold;

[0165] S32, when the initial tool wear level difference data set There is an initial tool wear level difference data greater than or equal to the wear level difference data threshold or initial energy consumption level data set When the energy consumption level data in the data is greater than or equal to the energy consumption level data threshold, the crop tillage depth dataset is satisfied. Under the premise of , the tillage depth data of the corresponding rotary tiller is adjusted until the initial tool wear level difference data set is There is no initial tool wear level difference data greater than or equal to the wear level difference data threshold and the initial energy consumption level data set The final tillage depth dataset is obtained until there is no energy consumption level data greater than or equal to the energy consumption level data threshold. represents the final tillage depth data of the i-th rotary tiller in the rotary tiller sample set;

[0166] Otherwise, there is no need to adjust the tillage depth data of the rotary tiller;

[0167] Adjusting the tillage depth data of the corresponding rotary tiller in S32 includes the following steps:

[0168] S321, constructing a first initial SVM classification model and a second initial SVM classification model;

[0169] S322, training and testing a first initial SVM classification model using the initial tillage depth dataset, the initial tool wear level difference dataset, the soil moisture average dataset, the soil hardness average dataset, and the soil texture level average dataset to obtain a first final SVM classification model;

[0170] The second initial SVM classification model is trained and tested using the initial tillage depth dataset, the initial energy consumption level dataset, the soil moisture average dataset, the soil hardness average dataset, and the soil texture level average dataset to obtain a second final SVM classification model;

[0171] In step S322, the first initial SVM classification model is trained and tested using the initial tillage depth dataset, the initial tool wear level difference dataset, the soil moisture average dataset, the soil hardness average dataset, and the soil texture level average dataset to obtain the first final SVM classification model, including the following steps:

[0172] S32211. Set a first training data ratio; divide the initial plowing depth dataset, the initial tool wear level difference dataset, the soil moisture average dataset, the soil hardness average dataset, and the soil texture level average dataset according to the first training data ratio to obtain a first initial plowing depth training dataset, an initial tool wear level difference training dataset, a first soil moisture average training dataset, a first soil hardness average training dataset, a first soil texture level average training dataset, a first initial plowing depth test dataset, an initial tool wear level difference test dataset, a first soil moisture average test dataset, a first soil hardness average test dataset, and a first soil texture level average test dataset;

[0173] S32212. Set a first training error threshold; input the first initial tillage depth training data set, the initial tool wear level difference training data set, the first soil moisture average training data set, the first soil hardness average training data set, and the first soil texture level average training data set into a first initial SVM classification model for training; during the training process, when the training error is less than the first training error threshold, stop training to obtain a first trained SVM classification model; otherwise, continue training until the training error is less than the first training error threshold;

[0174] S32213. Set a first test accuracy threshold; input the first initial plowing depth test data set, the initial tool wear level difference test data set, the first soil moisture average test data set, the first soil hardness average test data set, and the first soil texture level average test data set into the first trained SVM classification model for testing; after the test is completed, obtain a first test accuracy; when the first test accuracy is greater than or equal to the first test accuracy threshold, use the first trained SVM classification model as the first final SVM classification model; otherwise, return to S3212 to continue training the first trained SVM classification model until the first test accuracy is greater than or equal to the first test accuracy threshold;

[0175] In step S322, the second initial SVM classification model is trained and tested using the initial tillage depth dataset, the initial energy consumption level dataset, the soil moisture average dataset, the soil hardness average dataset, and the soil texture level average dataset to obtain the second final SVM classification model, including the following steps:

[0176] S32221. Set a second training data ratio; divide the initial plowing depth dataset, the initial energy consumption level dataset, the soil moisture average dataset, the soil hardness average dataset, and the soil texture level average dataset according to the second training data ratio to obtain a second initial plowing depth training dataset, an initial energy consumption level training dataset, a second soil moisture average training dataset, a second soil hardness average training dataset, a second soil texture level average training dataset, a second initial plowing depth test dataset, a second initial energy consumption level test dataset, a second soil moisture average test dataset, a second soil hardness average test dataset, and a second soil texture level average test dataset;

[0177] S32222. Set a second training error threshold; input the second initial tillage depth training data set, the initial energy consumption level training data set, the second soil moisture average training data set, the second soil hardness average training data set, and the second soil texture level average training data set into a second initial SVM classification model for training; during the training process, when the training error is less than the second training error threshold, stop training to obtain a second trained SVM classification model; otherwise, continue training until the training error is less than the second training error threshold;

[0178] S32223. Set a second test accuracy threshold; input the second initial plowing depth test data set, the initial energy consumption level test data set, the second soil moisture average test data set, the second soil hardness average test data set, and the second soil texture level average test data set into the second trained SVM classification model for testing; after the test is completed, obtain the second test accuracy; when the second test accuracy is greater than or equal to the second test accuracy threshold, use the second trained SVM classification model as the second final SVM classification model; otherwise, return to S3222 to continue training the second trained SVM classification model until the second test accuracy is greater than or equal to the second test accuracy threshold;

[0179] S323: Adjust the tillage depth data of the corresponding rotary tiller in conjunction with the first final SVM classification model and the second final SVM classification model to obtain a final tillage depth data set.

[0180] The S323 includes the following steps:

[0181] S3231: Set the tillage depth data to be adjusted and the corresponding tool wear level difference data f1′ to be optimized and the energy consumption level data f2′ to be optimized. Acquire the average soil moisture data, average soil hardness data, and average soil texture data corresponding to the tillage depth data to be adjusted by combining the average soil moisture data set, the average soil hardness data set, and the average soil texture data set. Record these data as average soil moisture adjustment data, average soil hardness adjustment data, and average soil texture adjustment data, respectively.

[0182] S3232, build the second squirrel population d2′ i represents the i-th squirrel in the second squirrel population, Represents the size of the second squirrel population; Set the maximum number of iterations of the second squirrel population to The current number of iterations is are respectively recorded as the second maximum number of iterations and the second current number of iterations; the search space dimension of the second squirrel population is 1 dimension;

[0183] S3233, setting the initial position of each squirrel in the second squirrel population according to the tillage depth data to be adjusted, and obtaining the initial position set f i represents the initial position of the i-th squirrel in the second squirrel population; the calculation formula is as follows,

[0184] f i =f min +rand i ·(f max -f min );

[0185] Where, f min 、f max Respectively represent the lower limit and upper limit of the tillage depth data to be adjusted, rand i For f i Generate a random number between 0 and 1;

[0186] S3234: Construct the fitness function of the second squirrel population based on the tool wear level difference data f1′ to be optimized and the energy consumption level data f2′ to be optimized. as follows,

[0187]

[0188] In the formula, β is a positive number, indicating the correction parameter;

[0189] S3235, start iteration, before iteration, set the second current iteration number Set to 1; the fitness function of the second squirrel population is used in the first iteration And cooperate with the first final SVM classification model and the second final SVM classification model to calculate the initial position set The fitness value of each initial position in the third fitness value set is obtained; the maximum fitness value and the corresponding initial position in the third fitness value set are respectively used as the third global optimal fitness and the third global optimal position; the initial position set is adjusted according to the third global optimal fitness and the third global optimal position. After the update is completed, the second current iteration number Add 1 and enter the next iteration;

[0190] The fitness function of the second squirrel population is used in each other iteration process And cooperate with the first final SVM classification model and the second final SVM classification model to calculate the fitness value of each position updated in the previous round of iteration to obtain a fourth fitness value set; the maximum fitness value in the fourth fitness value set and the corresponding position matrix are respectively used as the fourth global optimal fitness and the fourth global optimal position; each position updated in the previous round of iteration is continued to be updated according to the fourth global optimal fitness and the fourth global optimal position; after the update is completed, the second current iteration number is Add 1 and enter the next iteration;

[0191] S3236, when the When , the iteration is stopped and the second final global optimal position is output; the corresponding initial tillage depth data in the initial tillage depth data set is replaced by the second final global optimal position to obtain the final tillage depth data set

[0192] S4. adjusting parameters of rotary tillers corresponding to the plurality of cultivated fields according to the final tillage depth data set;

[0193] The S4 comprises the following steps:

[0194] S41, adjusting the parameters of each rotary tiller in the rotary tiller sample set to change the rotary tillage depth, and after the adjustment is completed, using ultrasound to actually measure the rotary tillage depth of each rotary tiller in the rotary tiller sample set to obtain an actual rotary tillage depth dataset;

[0195] S42. Set a rotary tillage depth error threshold; when the difference between the corresponding rotary tillage depth data in the actual rotary tillage depth data set and the final tillage depth data set is greater than or equal to the rotary tillage depth error threshold, return to S41 to continue parameter adjustment until the difference between the corresponding rotary tillage depth data in the actual rotary tillage depth data set and the final tillage depth data set is greater than or equal to the rotary tillage depth error threshold; otherwise, there is no need to return to S41 to continue parameter adjustment.

[0196] Example 2

[0197] This embodiment discloses a system for automatically controlling the tillage depth of a rotary tiller based on soil hardness detection. The system can implement the method of the above embodiment, including:

[0198] Throughout this specification, references to terms such as "one embodiment," "example," or "specific example" indicate that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the invention. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0199] The preferred embodiments of the invention disclosed above are intended only to help illustrate the invention. These preferred embodiments do not exhaust all details, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A method for automatically controlling the tillage depth of a rotary tiller based on soil hardness detection, characterized in that: The following steps are involved: S1. Collect average soil moisture data, average soil hardness data, and average soil texture grade data of multiple cultivated fields to obtain an average soil moisture dataset, an average soil hardness dataset, and an average soil texture grade dataset; And collect the tillage depth requirement data of crops corresponding to multiple cultivated lands to obtain the crop tillage depth data set; Said S1 comprises the following steps: S11, setting multiple plots of arable land of the same size that need to be rotary tilled and corresponding rotary tillers of the same type and quality to obtain a arable land sample set and a rotary tiller sample set; S12, measuring average soil moisture data, average soil hardness data, and average soil texture grade data of each cultivated land sample in the cultivated land sample set to obtain an average soil moisture data set, an average soil hardness data set, and an average soil texture grade data set; S13, setting tillage depth requirement data for crops planted in each piece of cultivated land sample in the cultivated land sample set to obtain a crop tillage depth dataset; S2. Calculating the tool wear level difference and energy consumption level data of rotary tillers corresponding to a plurality of cultivated fields before and after operation to obtain an initial tool wear level difference data set and an initial energy consumption level data set; The S2 comprises the following steps: S21. Setting a plurality of rotary tiller tool wear levels to obtain a tool wear level set; measuring tool wear level data of each rotary tiller in a rotary tiller sample set before operation based on the tool wear level set to obtain an initial tool wear level data set; setting initial tillage depth data based on the crop tillage depth data set to obtain an initial tillage depth data set; S22. Setting multiple energy consumption levels to obtain an energy consumption level set; measuring tool wear level data and energy consumption level data of each rotary tiller in the rotary tiller sample set according to the tool wear level set and the energy consumption level set to obtain a final tool wear level data set and an initial energy consumption level data set; subtracting the final tool wear level data set from the initial tool wear level data set to obtain an initial tool wear level difference data set; S3, using the initial tool wear level difference data set, the initial energy consumption level data set, the soil moisture average data set, the soil hardness average data set, and the soil texture level average data set to construct a first final SVM classification model and a second final SVM classification model; and optimizing and adjusting the initial tillage depth data of the rotary tillers corresponding to the plurality of cultivated fields in combination with the initial tool wear level difference data set, the initial energy consumption level data set, the soil moisture average data set, the soil hardness average data set, the soil texture level average data set, the first final SVM classification model, the second final SVM classification model, and the crop tillage depth data set to obtain a final tillage depth data set; The S3 includes the following steps: S31, setting a wear level difference data threshold and an energy consumption level data threshold; S32. When there exists in the initial tool wear degree level difference data set initial tool wear degree level difference data greater than or equal to the wear degree difference data threshold or there exists in the initial energy consumption level data set energy consumption level data greater than or equal to the energy consumption level data threshold, the plowing depth data of the corresponding rotary tiller is adjusted until there exists in the initial tool wear degree level difference data set initial tool wear degree level difference data greater than or equal to the wear degree difference data threshold and there exists in the initial energy consumption level data set energy consumption level data greater than or equal to the energy consumption level data threshold, thereby obtaining a final plowing depth data set; otherwise, there is no need to adjust the plowing depth data of the rotary tiller; Adjusting the tillage depth data of the corresponding rotary tiller in S32 includes the following steps: S321, constructing a first initial SVM classification model and a second initial SVM classification model; S322, using the initial tillage depth dataset, the initial tool wear level difference dataset, the soil moisture average dataset, the soil hardness average dataset, and the soil texture level average dataset to train and test a first initial SVM classification model to obtain a first final SVM classification model; using the initial tillage depth dataset, the initial energy consumption level dataset, the soil moisture average dataset, the soil hardness average dataset, and the soil texture level average dataset to train and test a second initial SVM classification model to obtain a second final SVM classification model; S323, adjusting the tillage depth data of the corresponding rotary tiller in conjunction with the first final SVM classification model and the second final SVM classification model to obtain a final tillage depth data set; S4. Adjust parameters of rotary tillers corresponding to the plurality of cultivated fields according to the final tillage depth data set.

2. The method for automatically controlling the tillage depth of a rotary tiller based on soil hardness detection according to claim 1, characterized in that: The S12 includes the following steps: S121, evenly setting multiple soil collection points on each cultivated land sample in the cultivated land sample set to obtain a soil collection point matrix; S122. Collect soil samples from corresponding cultivated land samples in the cultivated land sample set according to the soil collection point matrix to obtain a soil sample matrix; S123. Use a resistive soil moisture sensor to measure the moisture data of each soil sample in the soil sample matrix to obtain a soil moisture data matrix; use a soil hardness meter to measure the hardness data of each soil sample in the soil sample matrix to obtain a soil hardness data matrix; calculate the average value of each row of data in the soil moisture data matrix and the soil hardness data matrix to obtain a soil moisture average data set and a soil hardness average data set; measure the texture grade average data of each soil sample in each cultivated land sample in the soil sample matrix to obtain a soil texture grade average data set.

3. The method for automatically controlling the tillage depth of a rotary tiller based on soil hardness detection according to claim 2, characterized in that: The S323 includes the following steps: S3231: Set the tillage depth data to be adjusted and the corresponding tool wear level difference data to be optimized and the energy consumption level data to be optimized, and obtain the average soil moisture data, average soil hardness data, and average soil texture level data corresponding to the tillage depth data to be adjusted by combining the average soil moisture data set, the average soil hardness data set, and the average soil texture level data set, and record them as the average soil moisture adjustment data, the average soil hardness adjustment data, and the average soil texture level adjustment data, respectively. S3232, construct a second squirrel population; set the maximum number of iterations of the second squirrel population to , the current number of iterations is , respectively recorded as the second maximum number of iterations and the second current number of iterations; S3233, setting the initial position of each squirrel in the second squirrel population according to the tillage depth data to be adjusted, to obtain an initial position set; S3234: constructing a fitness function of the second squirrel population according to the tool wear level difference data to be optimized and the energy consumption level data to be optimized; S3235, start iteration; in each iteration, use the fitness function of the second squirrel population and cooperate with the first final SVM classification model, the second final SVM classification model, the soil moisture average adjustment data, the soil hardness average adjustment data, and the soil texture grade average adjustment data to calculate the fitness value of each position updated in the previous iteration, and continue to update each position updated in the previous iteration; S3236, when the When , the iteration is stopped and the second final global optimal position is output; the corresponding initial tillage depth data in the initial tillage depth dataset is replaced by the second final global optimal position to obtain the final tillage depth dataset.

4. The method for automatically controlling the tillage depth of a rotary tiller based on soil hardness detection according to claim 3 is characterized in that: The S4 comprises the following steps: S41, adjusting the parameters of each rotary tiller in the rotary tiller sample set to change the rotary tillage depth, and after the adjustment is completed, using ultrasound to actually measure the rotary tillage depth of each rotary tiller in the rotary tiller sample set to obtain an actual rotary tillage depth dataset; S42. Set a rotary tillage depth error threshold; when the difference between the corresponding rotary tillage depth data in the actual rotary tillage depth data set and the final tillage depth data set is greater than or equal to the rotary tillage depth error threshold, return to S41 to continue parameter adjustment until the difference between the corresponding rotary tillage depth data in the actual rotary tillage depth data set and the final tillage depth data set is greater than or equal to the rotary tillage depth error threshold; otherwise, there is no need to return to S41 to continue parameter adjustment.

5. A system for realizing the method for automatically controlling the tillage depth of a rotary tiller based on soil hardness detection as described in any one of claims 1 to 4.

Images