Rice and wheat harvester speed control method and control system
By obtaining the mechanical, image and environmental characteristics of the rice and wheat harvester in real time, combining multi-source data to predict the feeding amount and performing grading speed regulation, the lag problem of feeding amount control in traditional rice and wheat harvester is solved, and the operation efficiency and intelligence level are improved.
Patent Information
- Application Number
- CN202510914925.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-03
- Publication Date
- 2025-08-22
AI Technical Summary
The feeding volume control of traditional rice and wheat harvesters relies on manual experience, resulting in delayed response and low adjustment accuracy, making it difficult to meet the demands of precision agriculture for real-time and stability.
By obtaining the mechanical characteristics of the rice and wheat harvester, the characteristic images and environmental characteristics of the rice and wheat, the analysis of the images to obtain crop characteristics, combining multi-source data to predict the feeding amount in real time, and using the PID controller and comfort index function for grading speed regulation.
It realizes accurate prediction and real-time dynamic regulation of the feeding amount of rice and wheat harvester, improves operating efficiency and adaptability, and improves the intelligence level of the harvester.
Smart Images
Figure CN120524433A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of intelligent agricultural equipment, and in particular relates to a speed control method and control system for a rice and wheat harvester. Background Art
[0002] With the acceleration of agricultural mechanization and intelligentization, combine harvesters, as core equipment for efficient crop harvesting, have a significant impact on their performance, both in terms of efficiency and quality. The feed rate (i.e., the amount of crop entering the threshing and separation unit per unit time) is a key parameter in determining combine harvester performance. Excessive feed rates can lead to threshing drum blockage, increased grain breakage, and increased power loss; while insufficient feed rates reduce efficiency and waste resources. Therefore, accurately predicting and dynamically controlling feed rates is crucial for improving the adaptability, stability, and intelligence of combine harvesters.
[0003] Currently, feed rate control in traditional rice and wheat harvesters relies primarily on operator experience and judgment, with manual adjustments to the machine speed or header height indirectly adjusting feed rate. However, the complex and changing farmland operating environment presents challenges such as response lag and low adjustment accuracy, making it difficult to meet the real-time and stability requirements of precision agriculture. In recent years, some research has attempted to monitor feed rate using sensors (such as flow sensors and torque sensors) and provide feedback to the control system. However, this contact-based method requires the crop to enter the machine before it can take effect, resulting in a certain lag in achieving efficient and real-time control. Summary of the Invention
[0004] In view of the above analysis, the embodiments of the present invention aim to provide a speed control method and control system for a rice and wheat harvester, so as to solve the technical problem of hysteresis existing in the prior art.
[0005] The object of the present invention is achieved like this: A first embodiment of the present invention provides a rice and wheat harvester speed control method, comprising: Real-time acquisition of mechanical characteristics of rice and wheat harvesters, characteristic images of rice and wheat, and environmental characteristics; Analyze the characteristic images of rice and wheat to obtain crop characteristics including the number of wheat ears, crop height and crop grass-to-grain ratio; predicting a harvester feed rate based on the machine characteristics, crop characteristics, and environmental characteristics; Inputting the feed amount prediction value into the first speed control model and outputting the first speed control value; Inputting the first speed control value into a second speed control model and outputting a second speed control value; The speed of the harvester is controlled according to the second speed control value.
[0006] Furthermore, the mechanical characteristics include real-time speed, header height and grain moisture content; the environmental characteristics collect crops per unit area in the current harvesting environment through a five-point sampling method, and analyze the average weight, average height and average moisture content of the crops in the current harvesting environment.
[0007] Furthermore, the characteristic image of rice and wheat includes a front view characteristic image and a side view characteristic image. The front view characteristic image is obtained by collecting images in the forward direction of the harvester and calibrating them, and the side view characteristic image is obtained by collecting images on both sides perpendicular to the forward direction of the harvester and calibrating them.
[0008] Furthermore, the number of wheat ears is obtained by inputting the front view feature map into a pre-trained ear counting model. The pre-training process of the ear counting model is: collecting images of different environmental conditions, different speeds, and different angles in the forward direction of the harvester as a data set, dividing the data set into a training set and a test set, marking the number of wheat ears in all images in the training set, using the marked training set to iteratively train the ear counting model, and using the test set to verify the trained model until the accuracy meets the threshold requirement.
[0009] Furthermore, the crop height is calculated by inputting the side view feature map into a height detection model, segmenting the side view feature map to obtain a crop group image, calculating a first pixel difference between the highest point pixel value and the lowest point pixel value in the crop group image, and calculating the crop height based on a fitting function of the first pixel difference and the actual height.
[0010] Furthermore, the crop grass-to-grain ratio is obtained by inputting the side view feature map into a grass-to-grain ratio detection model, segmenting the side view feature map to obtain a plant group image, calculating a second pixel difference between the highest point pixel value and the lowest point pixel value in the plant group image, calculating the ear coefficient based on the first pixel difference and the second pixel difference, and calculating the crop grass-to-grain ratio based on a fitting function of the ear coefficient and the actual grass-to-grain ratio.
[0011] Furthermore, the feed amount of the harvester is predicted based on the machine characteristics, crop characteristics and environmental characteristics, which is expressed as: in, Indicates the predicted value of feed quantity, represents the number of wheat ears at time t, represents the real-time speed of the harvester at time t, 、 、 Respectively represent the average weight, average height and average moisture content of crops in the current harvesting environment, represents the moisture content of the grain at time t, represents the crop grass-to-grain ratio at time t, represents the crop height at time t, Indicates the height of the header at time t.
[0012] Furthermore, the step of inputting the feed amount prediction value into the first speed control model and outputting the first speed control value includes: The difference between the feed rate prediction value and the feed rate rating value is calculated, and the first speed control value is calculated based on the difference by the PID controller, which is expressed as: in, Indicates the rated value of feed volume. 、 and They represent the proportional coefficient, integral coefficient and differential coefficient respectively.
[0013] Furthermore, the first speed control value is input into the second speed control model, and the second speed control value is output, including: judging whether the value of the comfort index function satisfies the comfort range under the condition of the first speed control value; if so, the first speed control value is output as the second speed control value; otherwise, the parameters in the first speed control model are adjusted to generate a new first speed control value until the comfort range is satisfied, and the second speed control value is output.
[0014] A second embodiment of the present invention provides a rice and wheat harvester speed control system, comprising: A feature acquisition module is used to acquire the mechanical features of the rice and wheat harvester, the characteristic images of rice and wheat, and the environmental features in real time; A feature analysis module is used to analyze characteristic images of rice and wheat to obtain crop characteristics including the number of wheat ears, crop height, and crop grass-to-grain ratio; A feed amount prediction module, used for predicting the feed amount of the harvester based on the machine characteristics, crop characteristics and environmental characteristics; A first control module is used to input the feed amount prediction value into a first speed control model and output a first speed control value; a second control module, configured to input the first speed control value into a second speed control model and output a second speed control value; A speed control module is used to control the speed of the harvester according to the second speed control value.
[0015] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects: The rice and wheat harvester speed control method provided by the present invention detects the mechanical characteristics of the rice and wheat harvester, the characteristic images of rice and wheat, and the environmental characteristics in real time, then analyzes the characteristic images of rice and wheat to obtain crop characteristics, combines multi-source data such as mechanical characteristics, crop characteristics, and environmental characteristics to accurately predict the feed amount of the harvester in real time, and then performs graded speed control according to the feed amount prediction value. This method can not only meet the needs of real-time adjustment of the harvester speed, but also flexibly adapt to environmental conditions, thereby improving the working efficiency of the harvester. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] In order to more clearly illustrate the embodiments of this specification or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the embodiments of this specification. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.
[0017] Figure 1 Flowchart of the rice and wheat harvester speed control method provided in Example 1 of the present invention; Figure 2 Schematic diagram of the five-point sampling method provided in Example 1 of the present invention; Figure 3 Schematic diagram of front view feature map acquisition provided in Example 1 of the present invention; Figure 4 Schematic diagram of side view feature map acquisition provided in Example 1 of the present invention; Figure 5 This is a structural diagram of the speed control system of a rice and wheat harvester provided in Example 1 of the present invention. DETAILED DESCRIPTION
[0018] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. It should be noted that, in the absence of conflict, the embodiments and features in the embodiments disclosed in this disclosure can be combined, separated, interchanged and / or rearranged with each other. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0019] In the accompanying drawings, the sizes and relative sizes of components may be exaggerated for clarity and / or descriptive purposes. When the exemplary embodiments can be implemented differently, the specific process sequence may be performed in a different order than described. For example, two processes described in succession may be performed substantially simultaneously or in a reverse order from the described order. In addition, the same reference numerals represent the same components.
[0020] Example 1 A specific embodiment of the present invention, as Figure 1-4 As shown, a rice and wheat harvester speed control method is disclosed, comprising the following steps: S1. Real-time acquisition of mechanical features of rice and wheat harvesters, characteristic images of rice and wheat, and environmental features.
[0021] Specifically, the mechanical characteristics include real-time speed, header height, and grain moisture content; Figure 2 As shown, the environmental features are collected by a five-point sampling method for the crops per unit area in the current harvesting environment, and the average weight, average height and average moisture content of the crops in the current harvesting environment are analyzed; the real-time speed refers to the operating speed of the corn harvester, which can be obtained by a GNSS device, and the grain moisture content can be obtained by relevant sensors installed on the corn harvester; the characteristic images of rice and wheat include a front view characteristic image and a side view characteristic image, the front view characteristic image is obtained by collecting images in the forward direction of the harvester and calibrating them, and the side view characteristic image is obtained by collecting images on both sides perpendicular to the forward direction of the harvester and calibrating them; as shown Figure 3 and 4 As shown, cameras are installed on the front and side of the cab of the rice and wheat harvester, with a depression angle of 30°. Before operation, the images taken by the cameras are calibrated using a calibration plate. For example, the calibration area is a rectangle as shown in the figure, where = header length, , , for subsequent calculations.
[0022] S2. Analyze the characteristic images of rice and wheat to obtain crop characteristics including the number of wheat ears, crop height, and crop-grain ratio.
[0023] In this embodiment, the number of wheat ears is obtained by inputting the front view feature map into a pre-trained ear counting model. The pre-training process of the ear counting model is: collecting images of different environmental conditions, different speeds, and different angles in the forward direction of the harvester as a data set, dividing the data set into a training set and a test set, marking the number of wheat ears in all images in the training set, using the marked training set to iteratively train the ear counting model, and using the test set to verify the trained model until the accuracy meets the threshold requirement.
[0024] Exemplarily, the counting model can be a density map estimation model, and its main modules include three parts: a backbone feature extraction module, a pyramid multi-scale feature fusion module, and a centralized conversion module. The backbone feature extraction module adopts a lightweight Transformer model to improve the model's feature extraction advantages in terms of scale invariance and rotation invariance. The Transformer network is designed to have 4 layers, which respectively include shallow information and deep information. The pyramid multi-scale feature fusion module includes 4 layers of feature input, which are the features of the 4 layer outputs of the backbone network Transformer, and are effectively fused through channel splicing. By fusing the shallow detail features and deep semantic features of the image, the feature extraction of small target spikes is improved. The centralized conversion module is composed of channel attention, dynamic convolution feature changes, and spatial attention. By fusing the dynamic change features in the scene with the static basic features, such as dynamic changes under different lighting and different vehicle speeds, enhanced optimization of feature expression is achieved. In this embodiment, the crop height is obtained by inputting the side view feature map into a height detection model, segmenting the side view feature map to obtain a crop group image, calculating a first pixel difference between the highest point pixel value and the lowest point pixel value in the crop group image, and calculating the crop height according to a fitting function of the first pixel difference and the actual height; the crop grass-to-grain ratio is obtained by inputting the side view feature map into a grass-to-grain ratio detection model, segmenting the side view feature map to obtain a plant group image, calculating a second pixel difference between the highest point pixel value and the lowest point pixel value in the plant group image, calculating the ear coefficient according to the first pixel difference and the second pixel difference, and calculating the crop grass-to-grain ratio according to a fitting function of the ear coefficient and the actual grass-to-grain ratio.
[0025] For example, the height detection model collects different side view feature maps, first segments the crop group image from them, calculates the pixel difference between the highest point pixel value and the lowest point pixel value, and then performs regression fitting based on the actual height of the crop to obtain the relationship between the pixel difference and the plant height. ,in, Represents the first pixel difference; when used, input the first pixel difference and output the crop height; the grass-to-grain ratio detection model uses the same data set as the height detection model, first segmenting the plant group image to calculate the second pixel difference between the highest point pixel value and the lowest point pixel value, then dividing the first pixel difference and the second pixel difference to obtain the ear coefficient, and performing regression fitting based on the actual grass-to-grain ratio of the crop to obtain the relationship between the ear coefficient and the grass-to-grain ratio ,in, Represents the second pixel difference; when used, input the first pixel difference and the second pixel difference, and output the crop-grass-to-grain ratio.
[0026] S3. Predicting the feed amount of the harvester according to the machine characteristics, crop characteristics and environmental characteristics.
[0027] In this embodiment, step S3 is specifically expressed as follows: in, Indicates the predicted value of feed quantity, represents the number of wheat ears at time t, represents the real-time speed of the harvester at time t, 、 、 Respectively represent the average weight, average height and average moisture content of crops in the current harvesting environment, represents the moisture content of the grain at time t, represents the crop grass-to-grain ratio at time t, represents the crop height at time t, Indicates the height of the header at time t.
[0028] S4. Input the feed amount prediction value into the first speed control model and output the first speed control value.
[0029] Specifically, the working principle of the first speed control model is: The difference between the feed rate prediction value and the feed rate rating value is calculated, and the first speed control value is calculated based on the difference by the PID controller, which is expressed as: in, Indicates the rated value of feed volume. 、 and They represent the proportional coefficient, integral coefficient and differential coefficient respectively.
[0030] S5. Input the first speed control value into a second speed control model, and output a second speed control value.
[0031] Specifically, the first speed control value may not be the optimal solution, so it needs to be input into the second speed control model for verification. The verification process is as follows: Determine whether the value of the comfort index function satisfies the comfort range under the first speed control value condition; if so, output the first speed control value as the second speed control value; otherwise, adjust the parameters in the first speed control model to generate a new first speed control value until the comfort range is satisfied, and then output the second speed control value.
[0032] Exemplarily, the comfort index function is a function related to the mechanical vibration amplitude and acceleration, and is expressed as: ,in, Indicates the mechanical vibration amplitude, represents acceleration, Indicates the upper limit of mechanical vibration amplitude reference, Indicates the upper reference limit of acceleration, 、 Represents the weight coefficient, satisfying , the mechanical vibration amplitude can be obtained by the three-axis vibration acceleration sensor. , indicating that the current first speed control value satisfies the optimal solution, and it is directly output as the second speed control value to control the harvester speed, where Indicates the preset comfort threshold, which can be flexibly adjusted according to actual conditions. , indicating that the current first speed control value is not the optimal solution, and step S4 needs to be iterated again to obtain the second speed control value that satisfies the optimal solution, which is recorded as .
[0033] S6. Control the speed of the harvester according to the second speed control value.
[0034] Specifically, step S5 obtains the optimal speed control value After that, the operation speed controller adjusts the speed value according to the The voltage signal outputting the new speed adjusts the opening of the electromagnetic proportional valve, controls the flow of hydraulic oil to the hydraulic motor, and thus realizes the regulation of the harvester speed, which is expressed as: in, represents the target speed, Indicates the current operating speed. Indicates the second speed control value.
[0035] Compared with the existing technology, the rice and wheat harvester speed control method provided in this embodiment detects the mechanical characteristics of the rice and wheat harvester, the characteristic images of rice and wheat, and the environmental characteristics in real time, and then analyzes the characteristic images of rice and wheat to obtain crop characteristics. It combines multi-source data such as mechanical characteristics, crop characteristics and environmental characteristics to accurately predict the feed amount of the harvester in real time, and then performs graded speed control according to the feed amount prediction value. This can not only meet the real-time adjustment of the harvester speed, but also flexibly adapt to environmental conditions, thereby improving the working efficiency of the harvester.
[0036] Example 2 This embodiment provides a rice and wheat harvester speed control system. Figure 5 Shown, including: A feature acquisition module is used to acquire the mechanical features of the rice and wheat harvester, the characteristic images of rice and wheat, and the environmental features in real time; A feature analysis module is used to analyze characteristic images of rice and wheat to obtain crop characteristics including the number of wheat ears, crop height, and crop grass-to-grain ratio; A feed amount prediction module, used for predicting the feed amount of the harvester based on the machine characteristics, crop characteristics and environmental characteristics; A first control module is used to input the feed amount prediction value into a first speed control model and output a first speed control value; a second control module, configured to input the first speed control value into a second speed control model and output a second speed control value; A speed control module is used to control the speed of the harvester according to the second speed control value.
[0037] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A rice and wheat harvester speed control method, characterized in that: include: Real-time acquisition of mechanical characteristics of rice and wheat harvesters, characteristic images of rice and wheat, and environmental characteristics; Analyze the characteristic images of rice and wheat to obtain crop characteristics including the number of wheat ears, crop height and crop grass-to-grain ratio; predicting a harvester feed rate based on the machine characteristics, crop characteristics, and environmental characteristics; Inputting the feed amount prediction value into the first speed control model and outputting the first speed control value; Inputting the first speed control value into a second speed control model and outputting a second speed control value; The speed of the harvester is controlled according to the second speed control value.
2. The rice and wheat harvester speed control method according to claim 1, wherein: The mechanical characteristics include real-time speed, header height and grain moisture content; the environmental characteristics collect crops per unit area in the current harvesting environment through a five-point sampling method, and analyze the average weight, average height and average moisture content of the crops in the current harvesting environment.
3. The rice and wheat harvester speed control method according to claim 1, characterized in that: The characteristic images of rice and wheat include a front view characteristic image and a side view characteristic image. The front view characteristic image is obtained by collecting images in the forward direction of the harvester and calibrating them. The side view characteristic image is obtained by collecting images on both sides perpendicular to the forward direction of the harvester and calibrating them.
4. The rice and wheat harvester speed control method according to claim 3, characterized in that: The number of wheat ears is obtained by inputting the front view feature map into a pre-trained ear counting model. The pre-training process of the ear counting model is as follows: Images of the harvester under different environmental conditions, at different speeds, and at different angles in its forward direction are collected as a dataset. The dataset is divided into a training set and a test set. The number of wheat ears in all images in the training set is annotated. The ear counting model is iteratively trained using the annotated training set, and the trained model is verified using the test set until the accuracy meets the threshold requirement.
5. The rice and wheat harvester speed control method according to claim 3, characterized in that: The crop height is calculated by inputting the side view feature map into a height detection model, segmenting the side view feature map to obtain a crop group image, calculating a first pixel difference between the highest point pixel value and the lowest point pixel value in the crop group image, and calculating the crop height based on a fitting function of the first pixel difference and the actual height.
6. The rice and wheat harvester speed control method according to claim 5, characterized in that: The crop grass-to-grain ratio is obtained by inputting the side view feature map into a grass-to-grain ratio detection model, segmenting the side view feature map to obtain a plant group image, calculating a second pixel difference between the highest point pixel value and the lowest point pixel value in the plant group image, calculating the ear coefficient based on the first pixel difference and the second pixel difference, and calculating the crop grass-to-grain ratio based on a fitting function of the ear coefficient and the actual grass-to-grain ratio.
7. The rice and wheat harvester speed control method according to claim 6, characterized in that: The feed amount of the harvester is predicted based on the mechanical characteristics, crop characteristics and environmental characteristics, and is expressed as: in, Indicates the predicted value of feed quantity, represents the number of wheat ears at time t, represents the real-time speed of the harvester at time t, 、 、 Respectively represent the average weight, average height and average moisture content of crops in the current harvesting environment, represents the moisture content of the grain at time t, represents the crop grass-to-grain ratio at time t, represents the crop height at time t, Indicates the header height at time t.
8. The rice and wheat harvester speed control method according to claim 1, characterized in that: The step of inputting the feed amount prediction value into the first speed control model and outputting the first speed control value comprises: The difference between the feed rate prediction value and the feed rate rating value is calculated, and the first speed control value is calculated based on the difference by the PID controller, which is expressed as: in, Indicates the rated value of feed volume. 、 and They represent the proportional coefficient, integral coefficient and differential coefficient respectively.
9. The rice and wheat harvester speed control method according to claim 8, characterized in that: Inputting the first speed control value into a second speed control model and outputting a second speed control value includes: Determine whether the value of the comfort index function satisfies the comfort range under the first speed control value condition; if so, output the first speed control value as the second speed control value; otherwise, adjust the parameters in the first speed control model to generate a new first speed control value until the comfort range is satisfied, and then output the second speed control value.
10. A rice and wheat harvester speed control system, characterized in that: include: A feature acquisition module is used to acquire the mechanical features of the rice and wheat harvester, the characteristic images of rice and wheat, and the environmental features in real time; A feature analysis module is used to analyze characteristic images of rice and wheat to obtain crop characteristics including the number of wheat ears, crop height, and crop grass-to-grain ratio; A feed amount prediction module, used for predicting the feed amount of the harvester based on the machine characteristics, crop characteristics and environmental characteristics; A first control module is used to input the feed amount prediction value into a first speed control model and output a first speed control value; a second control module, configured to input the first speed control value into a second speed control model and output a second speed control value; A speed control module is used to control the speed of the harvester according to the second speed control value.