Combine harvester speed control method and control system

Abstract

The present invention relates to the technical field of agricultural equipment intelligence, and relates to a combine harvester speed control method and control system. The combine harvester speed control method comprises: acquiring the phenotypic characteristics of a crop and the operation speed and the working width of the harvester; inputting the phenotypic characteristics, the operation speed and the working width into a feed rate prediction model to output a feed rate prediction value; calculating a first speed control value on the basis of the feed rate prediction value and a feed rate rated value so as to adjust the speed of the harvester to a first speed; acquiring mechanical characteristics of the harvester after the crop enters the harvester; inputting the phenotypic characteristics and the mechanical characteristics into a feed rate detection model to output a feed rate accurate value, wherein the feed rate detection model is a multi-modal detection model obtained by means of data pre-training; and calculating a second speed control value by means of a speed fine-control model on the basis of the feed rate accurate value and the feed rate rated value so as to adjust the speed of the harvester to a second speed.

Description

A combine harvester speed control method and control system TECHNICAL FIELD

[0001] The present application belongs to the technical field of intelligent agricultural equipment, and in particular relates to a combine harvester speed control method and control system. BACKGROUND

[0002] It is urgent to develop agricultural production machinery with intelligence and modernization as the main body. As one of the most important agricultural harvesting machinery, the intelligent technology development level of the combine harvester to a great extent reflects the agricultural modernization degree and the agricultural mechanization development level of a country. Intelligent regulation and control of the combine harvester speed can significantly improve the operation efficiency of the harvester and reduce the labor intensity. The harvesting speed of most existing combine harvesters is still subjectively regulated by the driver according to the feeding amount. The driver with insufficient experience has difficulty in perceiving the overall load of the machine, which is prone to overload and underload working conditions, resulting in low harvesting efficiency.

[0003] Accurate detection of the feeding amount is a key index parameter for realizing operation speed control. The feeding amount refers to the amount of crops processed by the combine harvester per unit time. In actual field operation, the operation parameters need to be adjusted in real time to ensure the optimal feeding amount.

[0004] There are mainly two types of existing research methods for the feeding amount of the combine harvester: mechanical detection method, which detects relevant mechanical parameters according to sensors installed at different positions and fits the feeding amount through a regression model; and visual detection method, which identifies the phenotype characteristics of the crops to be harvested through image online recognition and then predicts the feeding amount. This method has prediction, but the existing image detection method mostly fits the feeding density at the next moment by recognizing the pixel value of the ear layer of the crops to be harvested per unit area through the camera; and has technical problems such as hysteresis and low accuracy. SUMMARY

[0005] In view of the above analysis, the embodiments of the present application aim to provide a combine harvester speed control method and control system to solve one or more of the above problems in the prior art.

[0006] The purpose of the present application is achieved as follows:

[0007] The first aspect of the present application provides a combine harvester speed control method, comprising:

[0008] Before the crops to be harvested enter the harvester, the phenotype characteristics of the crops and the operation speed and operation swath of the harvester are obtained;

[0009] The phenotype characteristics, the operation speed and the operation swath are input into a feeding amount prediction model to output a feeding amount prediction value at the next moment;

[0010] a first speed control value is calculated according to the feeding amount prediction value and the feeding amount rated value, so as to adjust the speed of the harvester to a first speed;

[0011] after the crops to be harvested enter the harvester, the mechanical characteristics of the harvester are obtained;

[0012] the phenotypic characteristics and the mechanical characteristics are input into a multi-modal feeding amount detection model to output a feeding amount accurate value;

[0013] a second speed control value is calculated according to the feeding amount accurate value and the feeding amount rated value, so as to adjust the speed of the harvester to a second speed.

[0014] Further, the phenotypic characteristics of the crops are obtained, including: collecting a phenotypic image of the crops by a depth camera; inputting the phenotypic image into a maturity detection model to detect the maturity period of the crops; the maturity period includes a milk maturity period, a wax maturity period, a full maturity period and a dry maturity period; and determining the water content of the crops according to the maturity period of the crops and the current time.

[0015] Further, the phenotypic characteristics of the crops are obtained, including: collecting a phenotypic image of the crops by a depth camera; inputting the phenotypic image into a maturity detection model to detect the maturity period of the crops; the maturity period includes a milk maturity period, a wax maturity period, a full maturity period and a dry maturity period; and determining the water content of the crops according to the maturity period of the crops and the current time.

[0016] Further, if the crops do not occur lodging, the phenotypic image is input into a density detection model to detect the density of the non-lodging crops; the density of the non-lodging crops includes four levels of low density, medium density, high density and extra-high density.

[0017] Further, the phenotypic characteristics of the crops are obtained, including: collecting a phenotypic image of the crops by a depth camera; inputting the phenotypic image into a maturity detection model to detect the maturity period of the crops; the maturity period includes a milk maturity period, a wax maturity period, a full maturity period and a dry maturity period; and determining the water content of the crops according to the maturity period of the crops and the current time.

[0018] Further, the mechanical characteristics include torque, rotational speed, pressure and vibration acceleration.

[0019] Further, the mechanical characteristics of the harvester are obtained, including: detecting the torque of the feeding auger driving shaft and the conveying bridge driving shaft by a torque sensor; detecting the rotational speed of the feeding auger driving shaft and the conveying bridge driving shaft by a rotational speed sensor; detecting the pressure of the conveying bridge bottom plate by a pressure sensor; and detecting the vibration acceleration of the conveying bridge by a three-axis gyroscope.

[0020] Further, if the crops occur lodging, the feeding amount prediction value is expressed as: Q pre=a1·W·L·V·H·S+a2

[0021] If the crop does not occur lodging, the feeding amount prediction value is expressed as: Q pre =a3·W·D·V·H·S+a4

[0022] Wherein, Q pre represents the feeding amount prediction value, a1, a2, a3 and a4 represent polynomial coefficients respectively, W represents the moisture content, L represents the density of the lodged crop, D represents the density of the non-lodged crop, V represents the working speed, H represents the plant height, and S represents the working swath.

[0023] Further, the input of the feeding amount detection model is the moisture content, the density of the lodged crop or the density of the non-lodged crop, the plant height, the torque of the feeding auger driving shaft, the rotating speed of the feeding auger driving shaft, the torque of the conveying bridge driving shaft, the rotating speed of the conveying bridge driving shaft, the pressure of the conveying bridge driving shaft and the vibration acceleration of the conveying bridge driving shaft, and the output is the feeding amount accurate value.

[0024] The second aspect embodiment of the present application provides a combine harvester speed control system, comprising:

[0025] A first acquisition module is configured to acquire the phenotypic characteristics of the crop and the working speed and working swath of the harvester.

[0026] A feeding amount prediction module is configured to input the phenotypic characteristics, the working speed and the working swath into a feeding amount prediction model to output a feeding amount prediction value.

[0027] A first speed adjustment module is configured to calculate a first speed control value according to the feeding amount prediction value and a feeding amount rated value, so as to adjust the speed of the harvester to a first speed.

[0028] A second acquisition module is configured to acquire the mechanical characteristics of the harvester after the crop enters the harvester.

[0029] A feeding amount detection module is configured to input the phenotypic characteristics and the mechanical characteristics into a feeding amount detection model to output a feeding amount accurate value; the feeding amount detection model is a multi-modal detection model obtained through data pre-training.

[0030] A second speed adjustment module is configured to calculate a second speed control value according to the feeding amount accurate value and the feeding amount rated value, so as to adjust the speed of the harvester to a second speed.

[0031] Compared with the prior art, the present application can at least achieve one of the following beneficial effects:

[0032] The combined harvester speed control method provided by the application predicts the feeding amount prediction value through the phenotypic characteristics of crops, pre-regulates the speed of the harvester according to the feeding amount prediction value, and then detects the feeding amount accurate value through the phenotypic characteristics and the mechanical characteristics of the harvester after the crops enter the harvester, fine-regulates the speed of the harvester according to the feeding amount accurate value, regulates the speed of the harvester through two stages, so that the feeding amount keeps the optimal value, eliminates the hysteresis of speed control, and ensures the accuracy of speed control. BRIEF DESCRIPTION OF DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present specification or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments described in the present specification, and other drawings can also be obtained by those skilled in the art according to these drawings.

[0034] Fig. 1 is a flow chart of the combined harvester speed control method provided by the embodiment 1 of the present application;

[0035] Fig. 2 is a flow chart of the feeding amount prediction provided by the embodiment 1 of the present application;

[0036] Fig. 3 is a flow chart of the pre-training of multiple detection models provided by the embodiment 1 of the present application;

[0037] Fig. 4 is a schematic diagram of the installation position of the depth camera provided by the embodiment 1 of the present application;

[0038] Fig. 5 is a schematic diagram of the structure of the feeding auger main shaft torque and speed integrated sensor provided by the embodiment 1 of the present application;

[0039] Fig. 6 is a schematic diagram of the structure of the conveying bridge main shaft torque and speed integrated sensor provided by the embodiment 1 of the present application;

[0040] Fig. 7 is a flow chart of detecting the feeding amount accurate value provided by the embodiment 1 of the present application;

[0041] Fig. 8 is a flow chart of speed adaptive regulation provided by the embodiment 1 of the present application;

[0042] Fig. 9 is a working schematic diagram of the harvester provided by the embodiment 1 of the present application;

[0043] Fig. 10 is a schematic diagram of the combined harvester speed control system provided by the embodiment 2 of the present application.

[0044] 1- chain wheel drive tooth; 2- first elastic body; 3- first rotor coil; 4- first stator coil; 5- first circuit board protection cover; 6- first hall speed sensor; 7- first magnet mounting seat; 8- shaft locking sleeve; 9- first stator fixing support; 10- first torque signal acquisition board; 11- belt pulley spoke; 12- second elastic body; 13- second stator coil; 14- second rotor coil; 15- second torque signal acquisition circuit board; 16- second circuit board protection cover; 17- second hall speed sensor; 18- second magnet mounting seat; 19- bearing end cover; 20- second stator fixing support. DETAILED DESCRIPTION

[0045] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. It should be noted that the embodiments and features in the present disclosure can be combined, separated, interchanged, and / or rearranged without conflict, if possible. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present application without creative labor fall within the protection scope of the present application.

[0046] In the drawings, the size and relative size of components can be exaggerated for clarity and / or descriptive purposes. When the exemplary embodiments can be carried out differently, a specific process sequence can be performed in an order different from that described. For example, two consecutively described processes can be performed substantially simultaneously or in an order opposite to that described. In addition, the same reference signs represent the same components.

[0047] Embodiment 1

[0048] One specific embodiment of the present application, as shown in FIGS. 1-2, discloses a combine harvester speed control method, comprising the following steps:

[0049] S1, obtaining the phenotypic characteristics of the crops and the working speed and working swath of the harvester.

[0050] In this embodiment, the phenotypic characteristics include moisture content, density, and plant height.

[0051] For example, to obtain the phenotypic characteristics, first, the phenotypic image and point cloud information of the crops are collected by a depth camera, and then the phenotypic characteristics are obtained from the phenotypic image and point cloud information, specifically comprising the following steps:

[0052] S101, input the phenotype image into a maturity stage detection model to detect the maturity stage of the crops; the maturity stage includes milk maturity stage, wax maturity stage, full maturity stage and dry maturity stage; and the water content of the crops is determined according to the maturity stage of the crops and the current time.

[0053] Specifically, different maturity stages have different water contents W at different times, and the water content values of the working period, generally 7-21 hours, are used to calibrate the water content of crops in different maturity stages; wherein the water content of the milk maturity stage at different times is W r7 , r8 , ... W r21 , the water content of the wax maturity stage at different times is W l7 , l8 , ... W l21 , the water content of the full maturity stage at different times is W w7 , w8 , ... W w21 , and the water content of the dry maturity stage at different times is W k7 , k8 , ... W k21 .

[0054] S102, input the phenotype image into a lodging detection model to detect whether the crops have lodged; if the crops have lodged, the density of the lodged crops is determined according to the lodging type and the lodging degree of the crops; the lodging type includes root lodging, stem lodging and stem folding; the lodging degree includes light lodging, medium lodging and heavy lodging; the density of the lodged crops is obtained by permutation and combination of the lodging type and the lodging degree; if the crops have not lodged, the density of the un-lodged crops is detected by inputting the phenotype image into a density detection model; the density of the un-lodged crops includes low density, medium density, high density and extra-high density.

[0055] Specifically, different lodging states correspond to different crop densities; when it is detected that the crops have lodged, the density needs to be determined according to the lodging type and the lodging degree of the crops, the lodging type and the lodging degree are obtained by image segmentation and then classification by the lodging detection model, the lodging type includes root lodging L g , stem lodging L j and stem folding L z , the lodging degree includes light lodging L l , medium lodging L m and heavy lodging L h , the lodging type and the lodging degree are combined to obtain the following 9 lodging states: {L g L l , L g L m , L g L hL j L l L j L m L j L h L z L l L z L m L z L h}

[0056] Corresponding to the density of the lodging crop respectively: {L1, L2, L3, L4, L5, L6, L7, L8, L9}

[0057] When it is detected that the crop does not occur lodging, it is necessary to classify the density of the non-lodging crop according to the crop quality in unit area by the density detection model, including low density D1∈(0<ρ<0.8), medium density D2∈(0.8≤ρ<1.5), high density D3∈(1.5≤ρ<2.2) and extra-high density D4∈(ρ≥2.2), wherein ρ represents the density of the crop in unit area, and the unit is kg / m 2 .

[0058] Exemplarily, the maturity detection model, the lodging detection model and the density detection model are built by a Transformer framework, and all the three detection models need to be pre-trained, and the specific process of pre-training is as follows:

[0059] As shown in FIG. 3, the data set is collected by using a vehicle-mounted depth camera and a drone, the data set is data-augmented to adapt to different harvesting environments, the data set is divided into a training set and a test set, 80% of which is used as the training set and 20% of which is used as the test set, the model is trained by using the divided training set, a pre-trained model is obtained, the pre-trained model is verified on the test set, an accuracy P is obtained, when the difference between the accuracy P and a true value P' is less than a preset threshold ε, the training phase is ended to obtain a detection model meeting the requirements, otherwise, the model is retrained until the requirements are met.

[0060] S103, collecting point cloud information of the crop by using a depth camera; inputting the point cloud information into the height detection model to detect the plant height.

[0061] Specifically, the specific working principle of this step is as follows:

[0062] The target crop is scanned using a depth camera to obtain three-dimensional point cloud data of the crop, and the depth camera captures detailed information of the crop surface, including height, width, shape and other three-dimensional features; the collected point cloud data is preprocessed, including denoising, filtering and segmentation operations, to ensure the accuracy and effectiveness of the data; the preprocessed point cloud data is input into a pre-trained height detection model, which calculates the height of the crop based on the input point cloud data, and the model identifies the distance between the highest point and the lowest point in the point cloud data and outputs the distance as the plant height.

[0063] For example, FIG. 4 is a schematic diagram of the installation angle of the depth camera. The installation position and angle of the depth camera will directly affect the style of the collected image, the image processing algorithm, and the accuracy of the feed rate prediction, etc. Assuming that the distance between the collected image and the combine harvester itself is x, the principle is to try to capture the near field information, x is related to the installation position h and the installation depression angle φ of the depth camera, f(x)=[h, φ], through the change of [h, φ], the distance between the captured crop information and the combine harvester can be obtained under the current view angle. The image acquisition effect is better when h∈[3m, 4m] and φ=30. When shooting the crop density dataset, a small red flag is set every 10m, and the crop density in the red flag interval is measured to match the corresponding position of the image.

[0064] It should be noted that the working speed can be obtained by the GNSS device, and the working swath can be determined by the model of the harvester or manually set, which is not limited in the embodiment.

[0065] S2, inputting the phenotype characteristics, the working speed and the working swath into a feed rate prediction model to output a feed rate prediction value.

[0066] Specifically, the feed rate at the next moment can be predicted according to the phenotype characteristics, the working speed and the working swath at the current moment, which can be divided into the following two cases:

[0067] (I) If the crop is lodged, the feed rate prediction value is represented as: Q pre =a1·W·L·V·H·S+a2

[0068] (II) If the crop is not lodged, the feed rate prediction value is represented as: Q pre =a3·W·D·V·H·S+a4

[0069] Wherein, Q pre represents the feed rate prediction value, a1, a2, a3 and a4 represent polynomial coefficients respectively, W represents the moisture content, L represents the density of the lodged crop, D represents the density of the non-lodged crop, V represents the working speed, H represents the plant height, and S represents the working swath.

[0070] It should be noted that the feed rate prediction model needs to be pre-trained before use. Specifically, the following steps are taken: phenotypic features of crop moisture content, lodging and non-lodging density, and plant height in the experimental area are collected using a five-point sampling method, and the corresponding operation speed and cutting width are obtained to form a dataset, which is then divided into a training set and a test set. The training set is used to train the feed rate prediction model, and the test set is used for verification. Training is stopped when the accuracy of the model is greater than a preset threshold, and a feed rate prediction model that meets the requirements is obtained.

[0071] S3. Calculate the first speed control value based on the predicted feed amount and the rated feed amount, so as to adjust the speed of the harvester to the first speed.

[0072] In this embodiment, after obtaining the predicted value of the feed amount in step S2, the first speed control value of the harvester speed is calculated based on the predicted value, and the harvester speed is adjusted to the first speed based on the value. Specifically, this includes the following steps:

[0073] S301. Input data, including the predicted feed amount Q. pre and feed rate rating Q rat In the input speed pre-regulation model.

[0074] Specifically, the feed rate rating Q rat Determined by the model of the harvester at the time of manufacture, the speed pre-regulation model can be a fuzzy adaptive PID control model.

[0075] S302. Difference Calculation: Calculate the predicted feed amount Q. pre and feed rate rating Q rat The first difference between them is expressed as follows: △Q1=Q pre -Q rat

[0076] Where △Q1 represents the first difference, Q pre Q represents the predicted feed rate. rat This indicates the rated feed rate.

[0077] S303. The first speed control value is calculated based on the first difference using fuzzy adaptive PID control, as shown below:

[0078] Where △V1 represents the first speed control value, △Q1 represents the first difference, and K p K represents the proportionality coefficient. i K represents the integral coefficient. d This represents the differential coefficient.

[0079] By dynamically adjusting the proportional coefficient K p Integral coefficient Ki and differential coefficient K d to obtain the optimized first speed regulation value AV1.

[0080] S304, adjusting the speed of the harvester to the first speed according to the first speed regulation value.

[0081] Specifically, as shown in FIG. 8, after obtaining the first speed regulation value AV1 in step S303, the opening of the electromagnetic proportional valve is adjusted by the voltage signal of the new speed output by the working speed controller according to the first speed regulation value AV1, the hydraulic oil flow to the hydraulic motor is controlled, and the pre-regulation of the speed of the harvester is realized, which is represented as follows: V1=V0+AV1

[0082] wherein V1 represents the first speed, V0 represents the current working speed, and AV1 represents the first speed regulation value.

[0083] S4, obtaining the mechanical characteristics of the harvester after the crops enter the harvester.

[0084] In the embodiment, the mechanical characteristics include torque, rotational speed, pressure, and vibration acceleration.

[0085] The obtaining of the mechanical characteristics specifically includes the following steps:

[0086] S401, detecting the torque of the feeding auger main shaft and the conveying bridge main shaft of the harvester by a torque sensor.

[0087] S402, detecting the rotational speed of the feeding auger main shaft and the conveying bridge main shaft by a rotational speed sensor.

[0088] For example, the torque sensor and the rotational speed sensor can be integrated, as shown in FIG. 5, which is a structural schematic diagram of the torque-rotational speed integrated sensor of the feeding auger main shaft, including a chain wheel transmission tooth 1, a first elastic body 2, a first rotor coil 3, a first stator coil 4, a first torque signal acquisition plate 10, a first circuit board protection cover 5, a first Hall speed detector 6, a first magnet mounting seat 7, a shaft locking sleeve 8, and a first stator fixed support 9.

[0089] The first elastic body 2 of the chain wheel structure is fixedly connected with the feeding auger driving shaft through a key, the chain drives the chain wheel transmission tooth 1 to rotate, thereby driving the driving shaft to rotate, the shaft locking sleeve 8 is used for locking the driving shaft and the first elastic body 2, preventing relative rotation, causing inaccurate torque measurement, the first elastic body 2 is deformed under the action of the load, the first torque signal acquisition plate 10 converts mechanical deformation into an electrical signal and performs AD conversion output, because the actual field harvesting environment is complex, the first circuit protection cover 5 is used for protecting the first torque signal acquisition plate 10;The first rotor coil 3 is located on the first elastic body 2, the first Hall speed sensor 6 is rotatably connected with the first rotor coil 3, the first stator coil 4 is kept at a certain distance from the first rotor coil 3 through a bearing, and is fixedly connected to the harvester shell through the first stator fixing support 9, the first magnet mounting seat 7 is installed on the first stator coil 4, and is fixed to the harvester shell together with the first stator coil 4, the first Hall speed sensor 6 calculates the rotating speed of the driving shaft by detecting the first magnet mounting seat 7 in the rotating process.

[0090] The structure diagram of the torque and speed integrated sensor of the conveying bridge driving shaft is shown in Figure 6, including the belt pulley spoke 11, the second elastic body 12, the second stator coil 13, the second rotor coil 14, the second torque signal acquisition circuit board 15, the second circuit board protection cover 16, the second Hall speed sensor 17, the second magnet mounting seat 18, the bearing end cover 19 and the second stator fixing support 20. The second elastic body 12 is fixedly connected with the conveying bridge driving shaft through a key, the belt drives the belt pulley spoke 11 to rotate, thereby driving the driving shaft to rotate, the bearing end cover 19 locks the driving shaft and the second elastic body 12 through a bolt, preventing relative rotation, causing inaccurate torque measurement, the second elastic body 12 is deformed under the action of the load, the second torque signal acquisition plate 15 converts mechanical deformation into an electrical signal and performs AD conversion output, because the actual field harvesting environment is complex, the second circuit board protection cover 16 is used for protecting the second torque signal acquisition plate 15;The second rotor coil 14 is installed on the second elastic body 12, the second Hall speed sensor 17 is rotatably connected with the second rotor coil 14, the second stator coil 13 is kept at a certain distance from the second rotor coil 14 through a bearing, and is fixed to the harvester shell structure through the second stator fixing support 20, the second magnet mounting seat 18 is installed on the second stator coil 13, and is fixed to the harvester shell structure together with the second stator coil 13, the second Hall speed sensor 17 calculates the rotating speed of the driving shaft by detecting the second magnet mounting seat 18 in the rotating process.

[0091] S403, detecting the pressure of the conveying bridge bottom plate through a pressure sensor.

[0092] Specifically, during the conveying operation, the pressure sensor collects the pressure data on the bottom plate in real time, converts the mechanical pressure into an electrical signal, and transmits it to the data acquisition system. The collected pressure data is preprocessed, including filtering, denoising, and calibration, to ensure the accuracy and reliability of the data, and then further analyzed to obtain the pressure value for predicting the feeding amount.

[0093] S404, detect the vibration acceleration of the conveying bridge by the three-axis gyroscope.

[0094] Specifically, during the conveying operation, the three-axis gyroscope collects vibration acceleration data of the conveying bridge in real time, converts the mechanical vibration into an electrical signal, and transmits it to the data acquisition system. The collected data is preprocessed, including filtering, denoising, and calibration, to ensure the accuracy and reliability of the data, and then further analyzed to obtain the vibration acceleration value for predicting the feeding amount.

[0095] S5, detecting the feeding amount accurate value by the feeding amount detection model based on the phenotype characteristics and the mechanical characteristics.

[0096] Specifically, after the speed of the harvester is adjusted to the first speed in step S3, the mechanical characteristics of the harvester are obtained by step S4 when the harvested crops enter the interior of the harvester, as shown in FIG. 7. Then the moisture content, the density of the lodged crops or the density of the non-lodged crops, the plant height, the torque of the feeding auger drive shaft, the rotational speed of the feeding auger drive shaft, the torque of the conveying bridge drive shaft, the rotational speed of the conveying bridge drive shaft, the pressure of the conveying bridge drive shaft, and the vibration acceleration of the conveying bridge drive shaft of the mechanical characteristics are input into the feeding amount detection model to detect the feeding amount accurate value, so as to output the feeding amount accurate value, which is represented as follows: Q act = f(W, L, H, T aug , ω aug , T con , ω con , P con , A con )

[0097] Alternatively, Q act = f(W, D, H, T aug , ω aug , T con , ω con , P con , A con )

[0098] wherein Q act represents the feeding amount accurate value, f(·) represents the feeding amount detection model, W represents the moisture content, L represents the lodged crop density, D represents the non-lodged crop density, H represents the plant height, T augrepresents the feeding auger drive shaft torque, ω aug represents the feeding auger drive shaft speed, T con represents the conveying bridge drive shaft torque, ω con represents the conveying bridge drive shaft speed, P con represents the conveying bridge bottom plate pressure, A con represents the conveying bridge vibration acceleration.

[0099] The feeding amount detection model is a multi-modal detection model obtained through data pre-training. The specific process is as follows: the phenotypic characteristics of the crop in the test area, such as moisture content, lodging and non-lodging density, and plant height, are collected by five-point sampling method; the mechanical characteristics of the feeding auger drive shaft torque, the feeding auger drive shaft speed, the conveying bridge drive shaft torque, the conveying bridge drive shaft speed, the conveying bridge bottom plate pressure, and the conveying bridge vibration acceleration are recorded synchronously by CAN bus technology; the actual value of the feeding amount is obtained by artificial receiving method; the above data is taken as a set of associated data set; more than 50 data sets are obtained; the data sets are divided into a training set and a test set; then a multi-sensor decision level fusion feeding amount detection model is established; the phenotypic characteristics and mechanical characteristics of the training set are taken as input, and the accurate value of the feeding amount is taken as output for training; the test set is used for verification; until the accuracy of the model is greater than a preset threshold, the training is stopped to obtain a feeding amount detection model meeting the requirements.

[0100] S6, a second speed control value is calculated according to the accurate value of the feeding amount and the rated value of the feeding amount, so as to adjust the speed of the harvester to a second speed.

[0101] In this embodiment, after obtaining the accurate value of the feeding amount in step S5, a second speed control value of the harvester speed is calculated according to the accurate value, and the speed of the harvester is adjusted to a second speed according to the value, as shown in FIG. 8. The specific steps include the following steps:

[0102] S601, input data, and input the accurate value Q act and the rated value Q rat of the feeding amount into the speed fine control model.

[0103] Specifically, the rated value Q rat of the feeding amount is determined by the model of the harvester when it is shipped, and the speed fine control model can also be a fuzzy self-adaptive PID control model.

[0104] S602, difference calculation, calculate the second difference between the accurate value Q act and the rated value Q rat of the feeding amount, which is represented as follows: ΔQ2=Q act -Q rat

[0105] wherein, ΔQ2 represents the second difference, Qact represents the feeding amount prediction value, Q rat represents the feeding amount rating value.

[0106] S603, calculating a second speed control value according to the second difference value through fuzzy adaptive PID control, represented as follows:

[0107] wherein, △V2 represents the second speed control value, △Q2 represents the first difference value, K' p represents the proportional coefficient, K' i represents the integral coefficient, K' d represents the differential coefficient.

[0108] By dynamically adjusting the proportional coefficient K' p , the integral coefficient K' i and the differential coefficient K' d , the optimal second speed control value △V2 is obtained.

[0109] S604, adjusting the speed of the harvester to a second speed according to the second speed control value.

[0110] Specifically, after obtaining the second speed control value △V2 in step S603, the voltage signal of the second speed is calculated by the working speed controller according to the second speed control value △V2 and the first speed, so as to fine-tune the opening of the electromagnetic proportional valve, control the hydraulic oil flow to the hydraulic motor, and further realize the fine control of the harvester speed, represented as follows: V2=V1+△V2

[0111] wherein, V2 represents the final speed, V1 represents the new speed after pre-control, and △V2 represents the second speed control value.

[0112] Compared with the prior art, the combined harvester speed control method provided in the embodiment predicts the feeding amount prediction value through the phenotypic characteristics of crops, pre-controls the speed of the harvester according to the feeding amount prediction value, accurately detects the feeding amount accurate value through the phenotypic characteristics and the mechanical characteristics of the harvester after the crops enter the inside of the harvester, fine-tunes the speed of the harvester according to the feeding amount accurate value, controls the speed of the harvester through two stages, keeps the feeding amount at the optimal value, eliminates the hysteresis of speed control, and ensures the accuracy of speed control.

[0113] Embodiment 2

[0114] The embodiment provides a combined harvester speed control system, as shown in FIG. 7, which comprises:

[0115] A first acquisition module is configured to acquire the phenotypic characteristics of crops and the working speed and working cutting width of the harvester.

[0116] The feeding amount prediction module is configured to input the phenotypic characteristics, the working speed, and the working swath into a feeding amount prediction model to output a feeding amount prediction value;

[0117] The first speed adjustment module is configured to calculate a first speed control value according to the feeding amount prediction value and a feeding amount rated value, so as to adjust the speed of the harvester to a first speed.

[0118] The second acquisition module is configured to acquire mechanical characteristics of the harvester after the crops enter the harvester.

[0119] The feeding amount detection module is configured to input the phenotypic characteristics and the mechanical characteristics into a feeding amount detection model to output a feeding amount accurate value; the feeding amount detection model is a multi-modal detection model obtained through data pre-training.

[0120] The second speed adjustment module is configured to calculate a second speed control value according to the feeding amount accurate value and the feeding amount rated value, so as to adjust the speed of the harvester to a second speed.

[0121] The above detailed description further describes the purpose, technical solutions, and beneficial effects of the present application. It should be understood that the above description is only a specific embodiment of the present application and is not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A speed control method for a combine harvester, characterized in that, include: To obtain the phenotypic characteristics of crops and the operating speed and cutting width of harvesters; The phenotypic features, the operating speed, and the operating width are input into the feed rate prediction model, which outputs the feed rate prediction value. The first speed control value is calculated based on the predicted feed amount and the rated feed amount to adjust the speed of the harvester to the first speed; Acquire the mechanical characteristics of the harvester after the crop enters it; The phenotypic features and mechanical features are input into the feed rate detection model, which outputs the precise feed rate value. The feeding amount detection model is a multimodal detection model obtained through data pre-training; The second speed control value is calculated based on the precise value and the rated value of the feed amount to adjust the speed of the harvester to the second speed.

2. The combine harvester speed control method according to claim 1, characterized in that, Obtaining crop phenotypic characteristics, including: Phenotypic images of crops are acquired using depth cameras; The phenotypic image is input into the maturity detection model to detect the maturity stage of the crop; The maturity period includes the milky stage, the waxy stage, the fully mature stage, and the withered stage; The moisture content of the crop is determined based on its maturity stage and the current time.

3. The combine harvester speed control method according to claim 2, characterized in that, Obtaining crop phenotypic characteristics also includes: The phenotypic image is input into the lodging detection model to detect whether the crop has lodged. If crops lodging occurs, the density of lodged crops should be determined based on the type and degree of lodging. The types of lodging include root lodging, stem lodging, and stem breakage; The degree of lodging includes mild lodging, moderate lodging, and severe lodging; The density of the lodged crops is obtained by arranging and combining the lodging type and the degree of lodging.

4. The combine harvester speed control method according to claim 3, characterized in that, If the crop has not lodged, the phenotypic image is input into the density detection model to detect the density of the non-lodged crop. The density of the non-lodged crops includes four levels: low density, medium density, high density, and extra-high density.

5. The combine harvester speed control method according to claim 4, characterized in that, Obtaining crop phenotypic characteristics also includes: Point cloud information of crops is collected using a depth camera; The point cloud information is input into the height detection model to detect the plant height.

6. The combine harvester speed control method according to claim 5, characterized in that, The mechanical characteristics include torque, rotational speed, pressure, and vibration acceleration.

7. The combine harvester speed control method according to claim 6, characterized in that, Obtain the mechanical characteristics of the harvester, including: The torque of the feed auger drive shaft and the conveyor bridge drive shaft of the harvester is detected by a torque sensor. The rotational speeds of the feed auger drive shaft and the conveyor bridge drive shaft are detected by speed sensors. The pressure of the conveyor bridge bottom plate is detected by a pressure sensor; The vibration acceleration of the conveyor bridge is detected by a three-axis gyroscope.

8. The combine harvester speed control method according to claim 4, characterized in that, If the crop lodging occurs, the predicted feed intake is expressed as follows: Q pre =a1·W·L·V·H·S+a2 If the crop does not lodging, the predicted feed intake is expressed as follows: Q pre =a3·W·D·V·H·S+a4 Among them, Q pre The value represents the predicted feed rate, a1, a2, a3, and a4 represent the polynomial coefficients, W represents the moisture content, L represents the density of lodged crops, D represents the density of unlodged crops, V represents the operating speed, H represents the plant height, and S represents the cutting width.

9. The combine harvester speed control method according to claim 7, characterized in that, The inputs to the feeding amount detection model are moisture content, density of lodged or non-lodged crops, plant height, torque of the feeding auger drive shaft, rotational speed of the feeding auger drive shaft, torque of the conveying bridge drive shaft, rotational speed of the conveying bridge drive shaft, pressure of the conveying bridge drive shaft, and vibration acceleration of the conveying bridge drive shaft. The output is the accurate value of the feeding amount.

10. A speed control system for a combine harvester, characterized in that, include: The first acquisition module is used to acquire the phenotypic characteristics of the crop and the operating speed and cutting width of the harvester; The feed rate prediction module is used to input the phenotypic features, the operating speed, and the operating width into the feed rate prediction model and output the feed rate prediction value. The first speed adjustment module is used to calculate the first speed control value based on the predicted feed amount and the rated feed amount, so as to adjust the speed of the harvester to the first speed. The second acquisition module is used to acquire the mechanical characteristics of the harvester after the crop enters the harvester; The feed amount detection module is used to input the phenotypic features and the mechanical features into the feed amount detection model and output the accurate feed amount value; the feed amount detection model is a multimodal detection model obtained through data pre-training; The second speed adjustment module is used to calculate a second speed control value based on the precise value of the feed amount and the rated value of the feed amount, so as to adjust the speed of the harvester to the second speed.

Images