A control method, terminal, and device for an electric power machine

Abstract

The application discloses a control method, terminal and device of an electric working machine, collects initial working ground information and calculates a target ploughing depth value; drives a cutter roll to reach the target ploughing depth value and performs working; collects multi-level ground image information after working of the cutter roll, adjusts the multi-level ground image information based on homography transformation, and calculates a current soil breaking rate according to the adjusted ground image information; and adjusts operation modes of a cutter roll motor and a walking motor according to the current soil breaking rate. The application continuously acquires multi-level ground image information after working, since the ground image information includes multiple angles for the same position, homography transformation is introduced, the multi-level ground image information is adjusted, the soil breaking rate achieved after the cutter roll attempts working is calculated according to the adjusted ground image information, and the operation modes of the cutter roll motor and the walking motor are adjusted based on the soil breaking rate, so that automatic control of the electric working machine during farming is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural machinery, and in particular to a control method, terminal and device for an electric working machine. Background Art

[0002] The most widely used small tillers in my country's hilly and mountainous areas are primarily fuel-powered micro-tillage machines and soil-raising machines. However, these fuel-powered small tillers lack green emissions, do not meet environmental requirements, and are not the future direction of development. In recent years, small electric tillers have appeared on the market. However, due to the heavy clay soil and harsh and highly variable working conditions in hilly and mountainous areas, these machines have poor overall performance and are not suitable for the heavy and frequently changing workloads. They require constant manual adjustment of tillage patterns and have even been associated with motor burnout due to overload. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a control method, terminal and device for an electric working machine to realize automatic control of the electric working machine during the farming process.

[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0005] A control method for an electric working machine, comprising the steps of:

[0006] S1. Collect initial ground information and calculate the target tillage depth;

[0007] S2, driving the cutter roller to reach the target tillage depth value and perform the operation;

[0008] S3, collecting multi-level ground image information after the cutter roller operation, adjusting the multi-level ground image information based on homography transformation, and calculating the current soil crushing rate based on the adjusted ground image information;

[0009] S4. Adjust the operation modes of the cutter roller motor and the travel motor according to the current soil crushing rate.

[0010] In order to solve the above technical problems, other technical solutions adopted by the present invention are:

[0011] A control terminal for an electric working machine comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps in the control method for an electric working machine are completed.

[0012] A device for executing the steps in the control method of an electric working machine, the device comprising a cutter roller motor, a travel motor and a frame, the cutter roller motor and the travel motor being assembled and connected to the frame, the cutter roller motor being used to turn over the soil on the working ground, and the travel motor being used to control the movement of the frame.

[0013] The beneficial effects of the present invention are: providing a control method, terminal and device for an electric working machine, calculating the target plowing depth value of the working ground and controlling the cutter roller to reach the target plowing depth value for trial operation, continuously obtaining multi-level ground image information after the operation during the trial operation, thereby obtaining the soil state, and because the ground image information includes multiple angles for the same position, a homography transformation is introduced to adjust the multi-level ground image information, and the soil crushing rate achieved after the cutter roller trial operation is calculated based on the adjusted ground image information, and the operating mode of the cutter roller motor and the travel motor is adjusted based on the soil crushing rate, thereby realizing automatic control of the electric working machine during farming. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 is a flow chart of a control method for an electric working machine in an embodiment of the present invention;

[0015] Figure 2 is a schematic diagram of a control terminal of an electric working machine in an embodiment of the present invention;

[0016] Description of labels:

[0017] 1. A control terminal for an electric working machine; 2. A memory; 3. A processor. DETAILED DESCRIPTION

[0018] To illustrate the technical content, achieved objectives and effects of the present invention in detail, the following description is given in conjunction with the embodiments and accompanying drawings.

[0019] Please refer to Figure 1 In an embodiment of the present invention, a method for controlling an electric working machine includes the steps of:

[0020] S1. Collect initial ground information and calculate the target tillage depth;

[0021] S2, driving the cutter roller to reach the target tillage depth value and perform the operation;

[0022] S3, collecting multi-layer ground image information after the cutter roller operation, adjusting the multi-layer ground image information based on homography transformation, and calculating the current soil crushing rate based on the adjusted ground image information;

[0023] S4. Adjust the operation modes of the cutter roller motor and the travel motor according to the current soil crushing rate.

[0024] A device for executing the steps in a control method of an electric working machine, the device comprising a cutter roller motor, a travel motor and a frame, the cutter roller motor and the travel motor being assembled and connected to the frame, the cutter roller motor being used to turn over the soil on the working ground, and the travel motor being used to control the movement of the frame.

[0025] It can be understood that the above-mentioned device is applied to the farming operation environment, and at least includes a frame body, a knife roller motor and a travel motor, wherein the knife roller motor and the travel motor work in coordination, that is, the travel motor controls the travel speed of the working machine, and the knife roller motor controls the working effect of the knife roller. The combination of the two ultimately achieves the crushing effect of the soil structure during micro-tillage or soil cultivation. However, since the soil crushing rate is inconvenient to detect, and the two operate relatively independently, a method for measuring the soil crushing rate and a method for automatically adjusting the operating modes of the two motors according to the soil crushing rate are proposed, which at least include the following beneficial effects:

[0026] By calculating the target tillage depth value of the working ground and controlling the cutter roller to reach the target tillage depth value for trial operation, multi-level ground image information after the operation is continuously obtained during the trial operation to obtain the soil state. Since the ground image information includes multiple angles for the same position, a homography transformation is introduced to adjust the multi-level ground image information. The soil crushing rate achieved after the cutter roller trial operation is calculated based on the adjusted ground image information, and the operating mode of the cutter roller motor and the travel motor is adjusted based on the soil crushing rate to realize the automatic control of the electric working machine during farming.

[0027] Specifically, the step S3 includes the following steps:

[0028] S31. Collect ground image information after the cutter roller operation, including a primary image and at least one auxiliary image. The ground images after the operation are captured using a camera, with the primary image representing the primary viewpoint and the auxiliary images representing additional images captured from different viewpoints or heights. These images together constitute basic data for subsequent soil condition analysis.

[0029] S32. Based on a homography, the feature information of the auxiliary image is converted to the main image to obtain a main feature map. The feature information of the auxiliary image is converted to the perspective of the main image using a homography algorithm. This conversion ensures that the information of the auxiliary image is consistent with that of the main image, ensuring the uniformity of the multi-view images, thereby generating a unified main feature map for subsequent processing.

[0030] S33. The main feature maps are stacked using a convolution operation to obtain a 3D cost volume, and soil block depth is measured on the 3D cost volume using a recursive multi-view stereo network (R-MVSNet) algorithm. The main image features are stacked using a convolution operation, and a 3D cost volume is formed by comprehensively considering complex factors such as occlusion, illumination changes, and perspective differences in soil blocks. This 3D cost volume contains image information from different perspectives, and each point contains image feature information from multiple perspectives. Backpropagation is used to automatically adjust parameters and automatically optimize the feature information in the 3D cost volume to reduce 3D reconstruction errors. Subsequently, the cost volume is analyzed using the recursive multi-view stereo network algorithm to obtain depth measurement information.

[0031] S34. Calculate the volume of the soil block and the current soil crushing rate using the microelement method using the image with depth measurement information.

[0032] Wherein, the step S32 specifically includes the steps of:

[0033] S321. Using a convolutional neural network to extract feature information of the main image and the auxiliary image; using a convolutional neural network to extract features of these images, it will carefully examine each photo, identify the outline and color features of the soil blocks in the photo, and use the difference in feature vector values ​​between the soil blocks and their surrounding debris to discard the surrounding debris of the soil blocks to improve the accuracy of subsequent three-dimensional reconstruction, thereby improving the accuracy of the soil crushing rate, and then use the homography matrix to transform the feature map of the auxiliary image to the corresponding depth plane under the perspective of the main image.

[0034] S322: Acquire built-in parameters of a visual module for collecting the ground image information, and convert the feature information of the auxiliary image into the main feature image using a homography transformation based on the built-in parameters.

[0035] The calculation formula of the homography transformation is expressed as:

[0036]

[0037] Where p i is the pixel coordinate system; d is the depth parameter of the camera; n T Represents the plane normal vector; I represents the world coordinate matrix of the camera; K1, R1, R1 T , C1, Z1 are the camera intrinsic matrix, rotation matrix, rotation matrix transpose, translation matrix, and depth of the main image respectively; K i 、R i 、C i 、Z i They are the camera intrinsic parameter matrix, rotation matrix, translation matrix, and depth of the auxiliary image respectively;

[0038] in, R1 -1and R i -1 is the matrix inverse operation, t1 is the value of the translation matrix of the main image, t i is the numerical value of the auxiliary graph's translation matrix.

[0039] The calculation principle of the above homography transformation is:

[0040] First, the built-in parameters of the visual module that collects the ground image information are obtained. The visual module can be a camera, and its built-in parameters include the intrinsic parameter matrix (K), the rotation matrix (R), the translation matrix (t), and the depth matrix (Z). These parameters describe the transformation relationship between the camera's world coordinate system and the camera coordinate system. Through the camera parameters, points in the world coordinate system can be transformed into the camera coordinate system, and further transformed into the pixel coordinate system. That is, the feature information of the captured pattern is transformed from the world coordinate system to the pixel coordinate system using the above parameters;

[0041] Next, a depth plane hypothesis is performed to calibrate the image's pixel coordinate system. Multiple depth planes are hypothesized in front of the main image, each with a corresponding depth value d. A depth plane is a plane hypothesized at a distance d from the camera lens.

[0042] Next, a homography is calculated. A homography describes the positional mapping between an object's world coordinate system and its pixel coordinate system. The corresponding transformation matrix is ​​called a homography. For a given depth value d, a homography matrix can be constructed from the auxiliary image to the primary image at depth d. This matrix transforms the feature points in the auxiliary image at depth d to their corresponding positions in the primary image's perspective.

[0043] Specifically, the step S33 includes the following steps:

[0044] S331. The main feature maps are stacked using a convolution operation to generate a 3D cost volume. Multiple convolution operations are performed to stack the feature information extracted from the main map into a 3D cost volume. The 3D cost volume is a composite of multi-view information, including feature information such as depth and position from different angles. This information will be used for depth estimation in subsequent steps.

[0045] S332. Using the Recursive Multi-View Stereo Network (R-MVSNet) algorithm, the probabilities of each soil block's depth measurement are output, and depth information is obtained using weighted averaging. By comparing these stacked features, R-MVSNet can infer the distance from each part of the photo to us, that is, the depth. It then uses reconstruction constraints across multiple images to predict the correct depth information.

[0046] Furthermore, the step S34 specifically includes the following steps:

[0047] Based on the Suzuki algorithm, the maximum distance, length, and volume of soil clods are calculated by restoring images with depth information. Based on the Suzuki algorithm, a pixel-by-pixel search and comparison method is used to separate the outer contours of soil clods. This effectively solves the problem of blurred boundaries due to adhesion of soil clods, improves the accuracy of soil clod separation, and thus improves the accuracy of soil crushing rate. Specifically, the Suzuki algorithm is used to search for the outer contours of adhered soil clods and then separate them. Finally, the OpenCV computer vision library uses the differential element method to restore the resulting image with depth information to calculate the maximum distance, length, and volume of soil clods.

[0048] Calculate the current soil crushing rate using the following formula:

[0049]

[0050] Where S is the soil crushing rate, Vi is the volume of soil blocks with a maximum distance length less than or equal to a preset length, and Vj is the volume of soil blocks with a maximum distance length greater than a preset length. Specifically, the preset length ranges from 3 to 5 cm, preferably 4 cm.

[0051] In addition, the step S1 further includes the following steps:

[0052] The calculation formula of the target tillage depth value is as follows:

[0053] H = L2 - L1 - δ;

[0054] Where H is the target tillage depth; L2 is the distance between the information collector and the lower quadrant of the cutter roller; L1 is the distance between the information collector and the ground; and δ is the wear radius of the cutter roller.

[0055] Specifically, in order to avoid frequent position adjustment of the knife roller, a buffer zone is set during the adjustment of the tillage depth value. The difference between the calculated tillage depth value H and the tillage depth value H1 set by the system is △H=H-H1; when △H≤0, the electric push rod controlling the knife roller retracts upward; when △H≤H2 (H2 is the tillage depth transition value, set by the system, for example, 2cm), the electric push rod does not move; when △H≥H2, the electric push rod extends downward.

[0056] Specifically, in the control method for the electric working machine of this embodiment, the combination of homography transformation and stacked convolution operations provides strong support for accurate soil crushing rate calculation and intelligent control of the working machine. First, a homography transformation is used to unify ground images captured from different perspectives and heights to the primary perspective, ensuring consistency between multi-perspective and multi-level image data. This step resolves the image inconsistency caused by perspective differences and lays a solid foundation for subsequent image processing and analysis. The system then uses convolution operations to extract features from these transformed images and stacks them, generating a three-dimensional cost volume that comprehensively considers complex factors such as soil clods due to occlusion, illumination changes, and perspective differences. This three-dimensional cost volume integrates and automatically optimizes image feature information from different dimensions using backpropagation to adjust parameters, enabling the system to more comprehensively and accurately describe the structure and state of the soil. Next, the system uses a recursive multi-view stereo network algorithm (R-MVSNet) to analyze the three-dimensional cost volume and accurately calculate the depth information of the soil clods. Based on this depth information, the volume of the soil clods is calculated using the microelement method, thereby determining the current soil crushing rate. This method, through multi-perspective, multi-level image fusion and precise depth measurement, not only improves the calculation accuracy of the soil fragmentation rate but also ensures real-time monitoring of the soil fragmentation status. Ultimately, the system intelligently adjusts the operating modes of the cutter roller motor and travel motor based on the real-time calculated soil fragmentation rate, enabling the electric work machine to automatically adapt to varying farmland soil conditions and optimize operational performance. This closed-loop control method not only improves operational efficiency but also achieves a higher level of intelligence and automation for electric work machines in the farming process.

[0057] Preferably, the step S4 further comprises the steps of:

[0058] Comparing the current soil crushing rate with the target soil crushing rate, and adjusting the operating power of the cutter roller motor and the travel motor based on the comparison result, specifically including:

[0059] Increasing or decreasing the speed of the knife roller motor;

[0060] and / or, increasing or decreasing the speed of the travel motor.

[0061] The operating power of the cutter roller motor and travel motor is adjusted mainly by controlling the speed, as follows:

[0062] Adjustment is performed based on the calculated soil crushing rate S and the preset working torque values ​​T1 and T2 of the travel motor and the cutter roller motor;

[0063] First, the system presets the target crushing rate gear (two gears of crushing rate, S1 and S2, S1 < S2), as well as the preset torque values ​​of the travel motor and the knife roller motor. The preset torque value of the travel motor is T 1a (Light load) and T1b (rated gear) two gears, T 1a <T 1b ; The preset torque value of the knife roller motor is T 2a (Light load) and T 2b (rated gear), T 2a <T 2b ; When working, adjust according to the following conditions:

[0064] (1) When S<S1,

[0065] ①T1≤T 1b And T2≤T 2a , N 2b =N 2a +△N2, where: N 2b The working speed of the knife roller motor after adjustment; N 2a is the current operating speed of the cutter roller motor; the initial value is set by the system, such as 2600r / min; △N2 is the cutter roller motor speed adjustment value, which is also set by the system, such as 100r / min; that is, when the travel motor is within the limit and the cutter roller motor is lightly loaded, the soil crushing rate can be improved by increasing the cutter roller motor speed (cutter roller speed);

[0066] ②T1≤T 1b And T 2a ≤T2≤T 2b , N 1b =N 1a -ΔN1, where: N 1b The working speed of the travel motor after adjustment; N 1a is the current operating speed of the travel motor, the initial value is set by the system, such as 2000r / min; ΔN1 is the travel motor speed adjustment value, which is also set by the system, such as 100r / min; that is, when the travel motor is within the limit (normal load or light load) and the cutter roller motor is under normal load, the soil crushing rate can be improved by reducing the travel motor speed (travel speed);

[0067] (2) When S1≤S≤S2, if T1<T 1a And T2<T 2a , N 1b =N 1a +ΔN1; that is, when the soil crushing rate is qualified, the travel motor is lightly loaded, and the cutter roller motor is lightly loaded, the operating efficiency can be improved by increasing the travel motor speed (travel speed);

[0068] (3) When S≥S2,

[0069] ①T1<T 1a And T2<T 2a , N 1b =N 1a+ΔN1; In this state, with a good soil crushing rate, light load on the travel motor and light load on the cutter roller motor, the operating efficiency can be improved by increasing the travel motor speed (travel speed);

[0070] ②T 1a ≤T1≤T 1b , N 2b =N 2a -△N2; that is, when the soil crushing rate is good and the travel motor is under normal load, the energy consumption is reduced by reducing the speed of the cutter roller motor (cutter roller speed) to achieve energy saving;

[0071] (4)T1>T 1b When N 1b =N 1a -ΔN1; that is, when the travel motor is overloaded, the travel motor load is reduced by reducing the travel motor speed (travel speed) to protect the travel motor;

[0072] (5)T2>T 2b When N 2b =N 2a -△N2, that is, when the cutter roller motor is overloaded, the cutter roller motor load is reduced by reducing the cutter roller motor speed (cutter roller speed) to protect the cutter roller motor; if S<S1, the system alarm will suspend the operation and request to modify the tillage depth and soil crushing rate values ​​set by the system.

[0073] The method of automatically adjusting the operating speed of the travel motor and the knife roller motor can realize the automatic, intelligent and continuous operation of the machine according to the soil crushing rate and load conditions of field operations, and has excellent operating efficiency, performance and energy saving.

[0074] Please refer to Figure 2 Another embodiment of the present invention provides a control terminal 1 for an electric working machine, comprising a memory 2, a processor 3, and a computer program stored in the memory 2 and executable on the processor 3. When the processor 3 executes the computer program, the steps of a method for controlling an electric working machine are performed. A terminal is provided as an execution medium for the method, configured to receive and process data and issue instructions for automated control.

[0075] In summary, the control method, terminal, and electric working machine provided in this application achieve precise control of the working process by accurately collecting working ground information, calculating the target tillage depth and soil crushing rate in real time, and combining dynamic adjustment of the cutter roller and travel motor. This has at least the following beneficial effects:

[0076] Real-time monitoring and feedback mechanism: By collecting multi-level ground image information after the cutter roller operation, using homography transformation and recursive multi-view stereo network algorithms, the soil condition after the operation can be monitored in real time. The operation results are fed back through the calculation of the soil crushing rate. This real-time monitoring and feedback mechanism ensures the immediate adjustment and optimization of the operation quality.

[0077] Automatically adjust operating modes: Based on a comparison of the current crushing rate with the target, the system automatically adjusts the operating modes of the cutter drum motor and travel motor. This feature allows the electric machine to dynamically adjust operating force and travel speed based on actual ground conditions, further optimizing operating efficiency and reducing manual intervention.

[0078] Intelligent Operation and Energy Saving: This invention avoids unnecessary energy consumption through intelligent adjustment of the cutter roller and travel motor. When the operating environment is ideal, the motor power is automatically reduced to reduce energy consumption. When soil conditions are complex, the machine can increase power to ensure the soil crushing rate meets the standard, thus achieving a balance between energy saving and high efficiency.

[0079] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent transformations made using the contents of the present invention's description and drawings, or directly or indirectly applied in related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A control method for an electric working machine, characterized in that: Including steps: S1. Collect initial ground information and calculate the target tillage depth; S2, driving the cutter roller to reach the target tillage depth value and perform the operation; S3, collecting multi-level ground image information after the cutter roller operation, adjusting the multi-level ground image information based on homography transformation, and calculating the current soil crushing rate based on the adjusted ground image information; S4. Adjusting the operating modes of the cutter roller motor and the travel motor according to the current soil crushing rate; The step S3 specifically includes the following steps: S31, collecting ground image information after the cutter roller operation, including a main image and at least one auxiliary image; S32, converting the feature information of the auxiliary image to the main image based on homography transformation to obtain a main feature image; S33, stacking the main feature maps using a convolution operation to obtain a three-dimensional cost volume, and using a recursive multi-view stereo network algorithm to measure the soil block depth on the three-dimensional cost volume; S34. Calculate the volume of the soil block and the current soil crushing rate using the image with depth measurement information through the microelement method.

2. The control method of an electric working machine according to claim 1, characterized in that: The step S32 specifically includes the following steps: S321, extracting feature information of the main image and the auxiliary image using a convolutional neural network; S322: Acquire built-in parameters of a visual module for collecting the ground image information, and convert the feature information of the auxiliary image into the main feature image using a homography transformation based on the built-in parameters.

3. The control method of an electric working machine according to claim 2, characterized in that: The calculation formula of the homography transformation is expressed as: Where p i is the pixel coordinate system; d is the depth parameter of the camera; n T Represents the plane normal vector; I represents the world coordinate matrix of the camera; K1, R1, R1 T , C1, Z1 are the camera intrinsic matrix, rotation matrix, rotation matrix transpose, translation matrix, and depth of the main image respectively; K i 、R i 、C i 、Z i They are the camera intrinsic parameter matrix, rotation matrix, translation matrix, and depth of the auxiliary image respectively; in R1 -1 and R i -1 is the matrix inverse operation, t1 is the value of the translation matrix of the main image, t i is the numerical value of the auxiliary graph's translation matrix.

4. The control method of an electric working machine according to claim 1, characterized in that: The step S33 specifically includes the following steps: S331, stacking the main feature maps using a convolution operation to obtain a three-dimensional cost volume; S332. Use a recursive multi-view stereo network algorithm to output the probability of the depth measurement value of each soil block, and use weighted average to obtain depth measurement information.

5. The control method of an electric working machine according to claim 1, characterized in that: The step S34 specifically includes the following steps: Based on the Suzuki algorithm, the maximum distance length and volume of soil blocks are calculated by using the microelement method to restore the image with depth measurement information. Calculate the current soil crushing rate using the following formula: Where S is the soil crushing rate, Vi is the volume of soil blocks with a maximum distance length less than or equal to the preset length, and Vj is the volume of soil blocks with a maximum distance length greater than the preset length.

6. The control method of an electric working machine according to claim 1, characterized in that: The step S1 further comprises the steps of: The calculation formula of the target tillage depth value is as follows: H = L2 - L1 - δ; Where H is the target tillage depth; L2 is the distance between the information collector and the lower quadrant of the cutter roller; L1 is the distance between the information collector and the ground; and δ is the wear radius of the cutter roller.

7. The control method of an electric working machine according to claim 1, characterized in that: The step S4 further comprises the steps of: Comparing the current soil crushing rate with the target soil crushing rate, and adjusting the operating power of the cutter roller motor and the travel motor based on the comparison result, specifically including: Increasing or decreasing the speed of the knife roller motor; and / or, increasing or decreasing the speed of the travel motor.

8. A control terminal for an electric working machine, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor completes the steps of the control method of an electric working machine as described in any one of claims 1 to 7 when executing the computer program.

9. A device, characterized in that: The device is used to execute the steps in the control method of an electric working machine described in any one of claims 1-7, and the device includes a knife roller motor, a travel motor and a frame, the knife roller motor and the travel motor are assembled and connected to the frame, the knife roller motor is used to turn over the soil on the working ground, and the travel motor is used to control the movement of the frame.

Images