Harvester header height control method and system based on multi-source sensing information fusion

Abstract

The invention discloses a harvester header height control method and system based on multi-source sensing information fusion, and belongs to the field of agricultural machinery automation control. The problems that in the prior art, actually-measured height detection deviation is large, and the height detection data source is single are solved, the longitudinal height and the transverse inclination angle of the header are calculated and adjusted in real time through multi-source sensor data, the accuracy of the height and the posture of the header is guaranteed from multiple dimensions so as to adapt to the complex working environment, and the working efficiency is improved. The precise control process of a hydraulic loop driving system on longitudinal and transverse hydraulic cylinders is defined, a reasonable height difference threshold value and a reasonable angle difference threshold value are set, fine regulation and control of the height of the header are achieved, abnormal conditions such as crop cutting missing and soil shoveling are reduced, crop and topographic feature points are extracted through a visual detection system and a deep convolutional neural network, and the accuracy of crop cutting is improved. A terrain curve and an operation plane can be accurately constructed, accurate operation path guidance is provided for a header, the accuracy and quality of harvesting operation are effectively improved, and the loss rate is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural machinery automation control, and in particular to a harvester header height control method and system based on multi-source sensor information fusion. Background Art

[0002] Harvester header profiling and height control technology has received widespread attention and development in recent years. Harvester header control is widely used in field crops such as rice, wheat, and soybeans. There are usually stubble height requirements when harvesting rice and wheat; the bottom pods of soybean plants are relatively low, and the header height needs to be precisely controlled to avoid improper header height control during the harvesting process, which may cause losses due to missing bottom pods or the header shoveling soil, affecting the quality and yield of soybean harvest.

[0003] Chinese patent publication number CN106068901A discloses a contouring harvester for harvesting tall and fallen crops, including a harvester frame, a harvester extension device, a feeding auger, a swing ring mechanism, a horizontal cutter, a reel support arm, a triangular reel and a contouring reel support device. The patent adopts a purely mechanical contouring mechanism, which relies on the gravity of the contouring mechanism to float and adhere to the ground. There is a hysteresis effect. When the crop density is high, the contouring mechanism float plate is easily lifted, and the actual measured height detection deviation is large.

[0004] Chinese patent publication number CN114616975B discloses an automatic profiling system for a combine harvester's cutting table and its control method, including a cutting table profiling device, a measurement system, an execution system, and a control system. The patent's height detection data comes from a single source and relies entirely on data from angle sensors installed on the profiling mechanism. Furthermore, the angle sensors are installed in a limited number of different forms, resulting in large errors in the constructed height detection model.

[0005] Therefore, it does not meet the existing needs. We propose a harvester cutting platform height control method and system based on multi-source sensor information fusion. Summary of the Invention

[0006] The purpose of the present invention is to provide a harvester cutting table height control method and system based on multi-source sensor information fusion, which calculates and adjusts the longitudinal height and lateral inclination of the cutting table in real time by utilizing multi-source sensor data, and ensures the accuracy of the cutting table height and posture in multiple dimensions. With the help of visual detection system and deep convolutional neural network to extract crop and terrain feature points, it can accurately construct terrain curves and working planes, effectively improve the accuracy and quality of harvesting operations, clarify the precise control process of the hydraulic circuit drive system on the longitudinal and lateral hydraulic cylinders, set reasonable height difference thresholds and angle difference thresholds, and realize fine regulation of the cutting table height, thereby solving the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solution: a harvester header height control method based on multi-source sensor information fusion, comprising the following steps:

[0008] The harvester is integrated with a header, a profiling mechanism, a first angle sensor, a second angle sensor, a transverse hydraulic cylinder, a conveying trough, a longitudinal hydraulic cylinder, a third angle sensor and a visual detection system;

[0009] The visual inspection system collects image information of crops to be harvested in real time, extracts the height feature points of crop bottom pods and the terrain feature points between crops and the ground, constructs the terrain curve and the optimal profiling operation plane, and determines the lateral rotation angle θ1 of the header and the height h of the soybean bottom pods. s ;

[0010] The real-time angle change differences Δθ1 and Δθ2 on both sides of the header are obtained through the profiling mechanism and the first angle sensor, and the real-time longitudinal height H of the header is calculated by combining the preset correlation coefficient and the correction coefficient;

[0011] The header obtains the lateral inclination angle change difference Δθ3 of the header through the second angle sensor to determine the lateral inclination angle of the header;

[0012] The longitudinal height adjustment value Δθ4 of the header is obtained through the conveyor trough and the third angle sensor. Combined with the height difference between the conveyor trough shaft and the frame and the distance from the center of the conveyor trough shaft to the cutting blade, the header height adjustment is verified to be in place.

[0013] According to the determined soybean bottom pod height h s The longitudinal target height H1 of the header is set based on the calculated real-time longitudinal height H of the header. When the difference between H and H1 exceeds the threshold, the longitudinal height of the header is adjusted by driving the longitudinal hydraulic cylinder.

[0014] According to the difference between the determined lateral rotation angle θ1 of the cutting platform and Δθ3 fed back by the second angle sensor, the lateral inclination angle of the cutting platform is adjusted by driving the lateral hydraulic cylinder to make the cutting platform match the terrain curve.

[0015] Furthermore, after the visual inspection system obtains the image of the soybean plant, it uses a deep convolutional neural network to extract the feature points of the soybean plant stem-pod connection and the feature points of the soybean plant and the ground connection. Based on the feature points of the soybean plant and the ground connection, points A and B are determined and a terrain curve between the soybeans to be harvested and the ground is constructed. At the same time, the optimal contouring operating plane of the harvesting platform is determined by connecting points A and B.

[0016] Furthermore, based on the characteristic points at the connection between the soybean plant stem and the bottom of the pod, the height difference h between each soybean plant and the line connecting points A and B is determined. s , the lifting height of the header cannot exceed h sThe angle θ1 between the line connecting points A and B and the horizontal line is the angle that the cutting table needs to rotate horizontally.

[0017] Furthermore, the calculation formula of the real-time longitudinal height H of the header is:

[0018]

[0019] Among them, k1 is the correlation coefficient; c1 is the correction coefficient.

[0020] Furthermore, the Δθ4 obtained by the third angle sensor is used to determine whether the header height is adjusted in place, and the calculation formula is as follows:

[0021]

[0022] Among them, H0 is the height difference between the conveyor trough shaft and the bottom of the harvester; L0 is the distance from the center of the conveyor trough shaft to the cutting knife.

[0023] The harvester header height control system based on multi-source sensor information fusion is applied to the harvester header height control method based on multi-source sensor information fusion, including:

[0024] A header is mounted on the harvester and is used to harvest crops, and the header is suspended on the conveyor trough through a hooking device;

[0025] The profiling mechanism is installed at the lower end of the cutting platform, with one set installed on each side of the cutting platform. One side of the profiling mechanism is hinged to the bottom of the cutting platform, and the other side is suspended on the cutting platform frame.

[0026] There are four angle sensors, namely the first angle sensor, the second angle sensor and the third angle sensor. The two first angle sensors are installed on the profiling mechanism on both sides of the cutting platform, the second angle sensor is installed for the lateral rotation of the cutting platform, and the third angle sensor is installed for the conveying trough. The angle sensors measure the real-time height and lateral inclination of the cutting platform and the rotation angle θ1 of the conveying trough;

[0027] The horizontal hydraulic cylinder allows the cutting platform to swing horizontally to adjust the height;

[0028] The longitudinal hydraulic cylinder is hingedly mounted on the harvester frame. The longitudinal hydraulic cylinder drives the conveyor trough to rotate, thereby raising and lowering the harvesting platform and changing the overall longitudinal height of the harvesting platform.

[0029] Among them, the horizontal hydraulic cylinder and the longitudinal hydraulic cylinder constitute a hydraulic circuit drive system, which includes two hydraulic circuits, one circuit realizes the overall longitudinal height control of the cutting platform, and the other circuit is used for the overall lateral height control of the cutting platform;

[0030] The visual inspection system, including a camera, a graphics acquisition card and an industrial computer, is used to collect images of crops to be harvested and analyze the crop height and soybean pod height information through image processing.

[0031] Furthermore, the hydraulic circuit drive system performs the following process:

[0032] The automatic control state of the cutting platform height is switched through the industrial computer. When the longitudinal height of the cutting platform is too low and the cutting platform is shoveling soil, DT1 is energized, oil enters the left end of the double-acting hydraulic cylinder, the piston rod is pushed out, and the cutting platform is lifted longitudinally as a whole;

[0033] When the header is too high and the bottom pods of soybeans are missed, the controller receives a signal to control the electromagnetic reversing valve, DT2 is energized, the right end of the double-acting hydraulic cylinder is filled with oil, the piston rod retracts, and the header descends longitudinally as a whole;

[0034] When the harvested land is uneven with the left side higher than the right, the controller controls DT3 to be energized, oil enters the left end of the double-acting hydraulic cylinder, the piston rod is pushed out, and the left side of the cutting platform is lifted to match the cutting platform with the ground shape;

[0035] When the harvested land is uneven with the right side higher and the left side lower, the controller controls DT4 to be energized, oil enters the right end of the double-acting hydraulic cylinder, the piston rod retracts, and the left side of the cutting platform falls back, so that the cutting platform matches the ground shape.

[0036] Furthermore, the overall longitudinal height of the header is regulated by performing the following process:

[0037] The visual inspection system collects soybean plant images, and the industrial computer detects and calculates h in real time. s and H1;

[0038] At the same time, the real-time longitudinal height H of the header is calculated based on Δθ1 and Δθ2 obtained from the first angle sensor data;

[0039] When |H-H1|≤10mm, the header and longitudinal hydraulic cylinder do not need to be adjusted;

[0040] When H-H1>10mm, the industrial computer outputs a signal control, the solenoid valve DT2 is energized, the right end of the double-acting hydraulic cylinder is filled with oil, the piston rod retracts, and the cutting platform is lowered longitudinally as a whole until |H-H1|≤10mm is satisfied;

[0041] When H1-H>10mm, the industrial computer outputs a signal control, the solenoid valve DT1 is energized, oil enters the left end of the double-acting hydraulic cylinder, the piston rod is pushed out, and the cutting table rises longitudinally as a whole until |H-H1|≤10mm is satisfied.

[0042] Furthermore, the overall lateral height of the header is regulated by performing the following process:

[0043] After the visual inspection system acquires the soybean plant image, it uses a deep convolutional neural network to extract the characteristic points where the soybean plant meets the ground. It then constructs the terrain curve between the soybeans to be harvested and the ground, as well as the line connecting points A and B, to determine the lateral rotation angle θ1 of the harvesting platform.

[0044] At the same time, according to the data of the second angle sensor, the lateral inclination angle change difference Δθ3 of the header is obtained;

[0045] When |θ1-Δθ3|≤2°, the header and the transverse hydraulic cylinder are not adjusted;

[0046] When θ1-Δθ3>2°, the controller controls DT3 to be energized, oil enters the left end of the double-acting hydraulic cylinder, the piston rod is pushed out, and the left side of the cutting platform is lifted, so that the cutting platform matches the ground shape until |θ1-Δθ3|≤2° is satisfied;

[0047] When Δθ3-θ1>2°, the controller controls DT4 to be energized, oil enters the right end of the double-acting hydraulic cylinder, the piston rod retracts, and the left side of the cutting platform falls back until |θ1-Δθ3|≤2° is satisfied, so that the cutting platform matches the ground shape.

[0048] Furthermore, the camera is used to obtain images of the morphology of the soybean plants to be harvested and their height from the ground, the graphics acquisition card is used to collect images, the industrial computer is used to process images and analyze crop height and soybean bottom pod height information, and the deep convolutional neural network is used to extract feature points at the connection between the soybean plant stem and pod and feature points at the connection between the soybean plant and the ground, to construct a terrain curve between the soybeans to be harvested and the ground and to determine the lateral rotation angle θ1 of the harvesting platform.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] 1. The present invention calculates and adjusts the longitudinal height and lateral inclination of the cutting platform in real time through multi-source sensor data, ensuring the accuracy of the cutting platform height and posture in multiple dimensions. Compared with single sensor control, the anti-interference ability is enhanced, the measurement and control are more reliable, and it can adapt to complex working environments.

[0051] 2. The present invention clarifies the precise control process of the longitudinal and transverse hydraulic cylinders by the hydraulic circuit drive system, sets reasonable height difference thresholds and angle difference thresholds, and realizes fine control of the cutting platform height, ensuring that the cutting platform operates stably at the target position during the harvesting process, reducing abnormal situations such as crop omission and soil shoveling, and improving work efficiency.

[0052] 3. The present invention extracts crop and terrain feature points through a visual detection system and a deep convolutional neural network, can accurately construct terrain curves and working planes, provide accurate working path guidance for the harvesting platform, and realize adaptive contour-profiling operations of the harvester's harvesting platform based on crop growth status and terrain, effectively improving the accuracy and quality of harvesting operations and reducing the loss rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a schematic diagram of the overall structure of the harvester of the present invention;

[0054] Figure 2 Schematic diagram of points A and B in the visual inspection system of the present invention;

[0055] Figure 3 A flow chart of the hydraulic circuit drive system of the present invention;

[0056] Figure 4 The figure is a flow chart of the longitudinal and lateral overall height control of the header of the present invention.

[0057] In the figure: 1. Cutting platform; 2. Profiling mechanism; 3. First angle sensor; 4. Second angle sensor; 5. Horizontal hydraulic cylinder; 6. Conveying trough; 7. Longitudinal hydraulic cylinder; 8. Third angle sensor; 9. Camera. DETAILED DESCRIPTION

[0058] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0059] In order to solve the technical problems that the existing technology uses a purely mechanical profiling mechanism, which relies on the gravity of the profiling mechanism to float and adhere to the ground, has a hysteresis effect, and is prone to causing the profiling mechanism float to rise when the crop density is high, resulting in large deviations in the measured height detection. At the same time, the height detection data source is single and completely relies on the data of the angle sensor installed on the profiling mechanism. The angle sensor installation form and number are limited, resulting in large errors in the constructed height detection model. Figure 1-Figure 4 , this embodiment provides the following technical solutions:

[0060] The method for controlling the height of a harvester header based on multi-source sensor information fusion includes the following steps:

[0061] The harvester is integrated with a header 1, a profiling mechanism 2, a first angle sensor 3, a second angle sensor 4, a transverse hydraulic cylinder 5, a conveying trough 6, a longitudinal hydraulic cylinder 7, a third angle sensor 8 and a visual detection system;

[0062] The visual inspection system collects image information of crops to be harvested in real time, extracts the height feature points of crop bottom pods and the terrain feature points between crops and the ground, constructs the terrain curve and the optimal profiling operation plane, and determines the lateral rotation angle θ1 of the header 1 and the height h of the soybean bottom pods. s, so that the harvester can operate according to the actual crop growth conditions and terrain characteristics, effectively avoiding problems such as incomplete harvesting or crop damage caused by the header 1 being too high or too low, thereby improving the harvesting quality and reducing crop losses;

[0063] The real-time angular difference values Δθ1 and Δθ2 on both sides of the header 1 are obtained through the profiling mechanism 2 and the first angle sensor 3. The real-time longitudinal height H of the header 1 is calculated by combining the preset correlation coefficient and the correction coefficient. This process ensures that the header 1 always operates at an appropriate height, closely following the growth height of the crops and terrain changes, further reducing missed harvests and losses during the harvesting process, ensuring the consistency and stability of the harvesting effect, and enhancing the harvester's adaptability to different terrains and crop growth conditions.

[0064] The header 1 obtains the lateral inclination angle difference Δθ3 of the header 1 through the second angle sensor 4 to determine the lateral inclination angle of the header 1. The lateral inclination angle control mechanism can prevent uneven harvesting, crop accumulation or damage caused by improper lateral inclination of the header 1 during operation, thereby improving the stability and harvesting quality of the harvester when operating on complex terrain, enhancing the overall operating performance, and reducing the impact of undulating or uneven terrain on the harvesting effect;

[0065] The longitudinal height adjustment amount Δθ4 of the header 1 is obtained through the conveyor trough 6 and the third angle sensor 8. Combined with the height difference between the conveyor trough 6 rotating shaft and the frame and the distance from the center of the conveyor trough 6 rotating shaft to the cutting blade, it is verified whether the height adjustment of the header 1 is in place. This verification step ensures the accuracy of the longitudinal hydraulic cylinder 7 in adjusting the height of the header 1, avoids harvesting problems caused by inadequate height adjustment, improves the operating accuracy and reliability of the harvester, and provides a strong guarantee for high-quality harvesting operations.

[0066] According to the determined soybean bottom pod height h s and the calculated real-time longitudinal height H of the cutting platform 1, setting the longitudinal target height H1 of the cutting platform 1, and when the difference between H and H1 exceeds the threshold, adjusting the longitudinal height of the cutting platform 1 by driving the longitudinal hydraulic cylinder 7;

[0067] According to the difference between the determined lateral rotation angle θ1 of the cutting platform 1 and the Δθ3 fed back by the second angle sensor 4, the lateral inclination angle of the cutting platform 1 is adjusted by driving the lateral hydraulic cylinder 5 so that the cutting platform 1 matches the terrain curve;

[0068] The calculation formula for the longitudinal real-time height H of the header 1 is:

[0069]

[0070] Among them, k1 is the correlation coefficient; c1 is the correction coefficient.

[0071] In this embodiment, k1=73.5, c1=3 mm, and the unit of H is mm;

[0072] The Δθ4 obtained by the third angle sensor 8 is used to determine whether the height of the header 1 is adjusted in place. The calculation formula is as follows:

[0073]

[0074] Among them, H0 is the height difference between the rotating shaft of the conveyor trough 6 and the bottom of the harvester; L0 is the distance from the center of the rotating shaft of the conveyor trough 6 to the cutting knife; it should be noted that H0 and L0 are both fixed values for different harvesters.

[0075] The technical effect of the above technical solution is: the comprehensive use of multi-source sensor information such as the visual detection system, multiple angle sensors, and the horizontal hydraulic cylinder 5 and the longitudinal hydraulic cylinder 7 realizes multi-dimensional, real-time and precise control of the height of the harvester's cutting table 1. The multi-source information fusion method improves the intelligence level and automation level of the harvester, reduces manual intervention, reduces the difficulty of operation and labor intensity, improves the operation efficiency and quality stability, enables the harvester to better adapt to the large-scale, efficient and precise harvesting needs of modern agriculture, and promotes the development and progress of agricultural harvesting technology.

[0076] After the visual inspection system obtains the image of the soybean plant, it uses a deep convolutional neural network to extract the feature points of the soybean plant stem-pod connection and the feature points of the soybean plant-ground connection. Based on the feature points of the soybean plant-ground connection, points A and B are determined and the terrain curve between the soybeans to be harvested and the ground is constructed. At the same time, the optimal profiling operating plane of the harvesting platform 1 is determined by connecting points A and B.

[0077] The technical effect of the above technical solution is: the use of deep convolutional neural network, an advanced image processing technology, realizes the efficient and accurate extraction of characteristic points of soybean plants, improves the intelligence level and automation level of the harvester, and constructs the terrain curve between the soybeans to be harvested and the ground based on the A and B points determined by the characteristic points at the connection between the soybean plant and the ground, providing the harvester with an accurate terrain reference, so that the harvester can operate according to the actual terrain. At the same time, the optimal contour-matching operating plane of the cutting platform 1 is determined by connecting the two points A and B, ensuring that the cutting platform 1 can stick to the terrain curve during operation, effectively avoiding problems such as incomplete harvesting or the cutting platform 1 colliding with the ground due to undulating terrain, improving the harvesting quality and efficiency, and reducing crop losses and mechanical damage.

[0078] Based on the characteristic points at the connection between the soybean plant stem and the bottom of the pod, determine the height difference h between each soybean plant and the line connecting points A and B. s , the lifting height of the header 1 cannot exceed h s The angle θ1 between the line connecting points A and B and the horizontal line is the angle that the cutting platform 1 needs to rotate horizontally.

[0079] The technical effect of the above technical solution is: according to the characteristic points of the soybean plant stem-pod bottom connection, the height difference h between each soybean plant and the line connecting points A and B is determined. s , clearly state that the lifting height of the header 1 cannot exceed h s , thereby ensuring that the height of the header 1 can be accurately controlled during the harvesting process, avoiding unnecessary damage to the soybean plants caused by the header 1 being too high or too low, and ensuring the integrity and efficiency of the harvesting operation. The lateral rotation angle of the header 1 is determined by the angle θ1 between the line connecting points A and B and the horizontal line, so that the header 1 can flexibly adjust the lateral inclination angle according to terrain changes, better fit the growth status of the soybean plants and terrain characteristics, further improve the harvesting effect, and reduce problems such as uneven harvesting caused by improper lateral inclination angle.

[0080] The harvester header height control system based on multi-source sensor information fusion is applied to the harvester header height control method based on multi-source sensor information fusion, including:

[0081] The header 1 is mounted on the harvester and is used to harvest crops, and the header 1 is suspended on the conveying trough 6 through a hooking device;

[0082] The profiling mechanism 2 is installed at the lower end of the cutting platform 1, with one set installed on each side of the cutting platform 1. One side of the profiling mechanism 2 is hinged to the bottom of the cutting platform 1, and one side is suspended on the frame of the cutting platform 1.

[0083] There are four angle sensors, namely the first angle sensor 3, the second angle sensor 4 and the third angle sensor 8. The two first angle sensors 3 are provided on the profiling mechanism 2 on both sides of the cutting platform 1, the second angle sensor 4 is provided for the lateral rotation of the cutting platform 1, and the third angle sensor 8 is provided for the conveying trough 6. The angle sensors measure the real-time height and lateral inclination of the cutting platform 1 and the rotation angle θ1 of the conveying trough 6;

[0084] The horizontal hydraulic cylinder 5 makes the cutting platform 1 swing horizontally to adjust the height;

[0085] The longitudinal hydraulic cylinder 7 is hingedly mounted on the harvester frame, and drives the conveying trough 6 to rotate, thereby raising and lowering the cutting platform 1 and changing the overall longitudinal height of the cutting platform 1;

[0086] Among them, the horizontal hydraulic cylinder 5 and the longitudinal hydraulic cylinder 7 constitute a hydraulic circuit drive system, which includes two hydraulic circuits, one circuit realizes the longitudinal height control of the cutting platform 1, and the other circuit is used for the lateral height control of the cutting platform 1;

[0087] The visual inspection system includes a camera 9, a graphics acquisition card, and an industrial computer, which is used to collect images of crops to be harvested and analyze crop height and soybean pod height information through image processing, wherein:

[0088] Camera 9 is used to obtain images of the morphology of the soybean plants to be harvested and their height relative to the ground. The graphics acquisition card is used to collect images. The industrial computer is used to process images and analyze the crop height and soybean bottom pod height information. The feature points of the soybean plant stem-pod connection and the feature points of the soybean plant-ground connection are extracted through a deep convolutional neural network to construct a terrain curve between the soybeans to be harvested and the ground and determine the lateral rotation angle θ1 of the harvesting platform 1.

[0089] The technical effect of the above technical solution is: through four angle sensors, various posture information of the cutting platform 1 is measured in real time, including real-time height, lateral inclination and conveyor trough rotation angle, etc., to provide accurate data basis for the height control of the cutting platform 1. At the same time, combined with the analysis of crop height and soybean bottom pod height information by the visual detection system, the longitudinal and lateral heights of the cutting platform can be accurately controlled to avoid damage to soybean plants caused by cutting 1 too high or too low, thereby ensuring the integrity and efficiency of the harvesting operation. The hydraulic circuit drive system realizes the longitudinal height control and lateral height regulation of the cutting platform 1 respectively, and cooperates with the profiling mechanism 2 to enable the cutting platform 1 to flexibly adapt to different terrain conditions, closely fit the ground and crop growth status, and maintain a stable working posture even in uneven fields, ensuring the uniformity and consistency of harvesting, and enhancing the stability and adaptability of the harvester when operating in complex terrain.

[0090] The hydraulic circuit drive system performs the following processes:

[0091] The automatic height control state of the cutting platform 1 is switched through the industrial computer. When the longitudinal height of the cutting platform 1 is too low and the cutting platform 1 is shoveling soil, DT1 is energized, oil enters the left end of the double-acting hydraulic cylinder, the piston rod is pushed out, and the cutting platform 1 is lifted longitudinally as a whole;

[0092] When the height of the cutting platform 1 is too high and the bottom pods of soybeans are missed, the controller receives the signal to control the electromagnetic reversing valve, DT2 is energized, the right end of the double-acting hydraulic cylinder is filled with oil, the piston rod retracts, and the cutting platform 1 is lowered longitudinally as a whole;

[0093] When the harvested land is uneven with the left side higher than the right, the controller controls DT3 to be energized, oil is supplied to the left end of the double-acting hydraulic cylinder, the piston rod is pushed out, and the left side of the cutting platform 1 is lifted, so that the cutting platform 1 matches the ground shape;

[0094] When the harvested land is uneven with the right side higher and the left side lower, the controller controls DT4 to be energized, oil enters the right end of the double-acting hydraulic cylinder, the piston rod retracts, and the left side of the cutting platform 1 falls back, so that the cutting platform 1 matches the ground shape.

[0095] The technical effects of the above technical solution are: Figure 3As shown, the hydraulic circuit drive system can quickly respond to the needs of adjusting the cutting platform posture under different terrain conditions. By accurately controlling the telescopic action of the hydraulic cylinder, the height and inclination of the cutting platform 1 can be smoothly adjusted, avoiding violent shaking or sudden changes of the cutting platform 1 during operation, thereby improving the overall operation stability of the harvester. The entire control process is automated, from the monitoring and judgment of the height and inclination to the execution of the hydraulic cylinder action, without the need for manual operation of the hydraulic system, which greatly improves the degree of automation of the harvesting operation.

[0096] To adjust the overall longitudinal height of the header 1, perform the following steps:

[0097] The visual inspection system collects soybean plant images, and the industrial computer detects and calculates h in real time. s and H1;

[0098] At the same time, the real-time longitudinal height H of the header 1 is calculated based on the Δθ1 and Δθ2 obtained from the first angle sensor 3;

[0099] When |H-H1|≤10mm, the header 1 and the longitudinal hydraulic cylinder 7 are not adjusted;

[0100] When H-H1>10mm, the industrial computer outputs a signal control, the solenoid valve DT2 is energized, the right end of the double-acting hydraulic cylinder is filled with oil, the piston rod retracts, and the cutting platform 1 is lowered longitudinally as a whole until |H-H1|≤10mm is satisfied;

[0101] When H1-H>10mm, the industrial computer outputs a signal control, the solenoid valve DT1 is energized, oil enters the left end of the double-acting hydraulic cylinder, the piston rod is pushed out, and the cutting platform 1 rises longitudinally as a whole until |H-H1|≤10mm is satisfied.

[0102] To adjust the overall lateral height of the header 1, perform the following steps:

[0103] After the visual inspection system acquires the soybean plant image, it uses a deep convolutional neural network to extract the characteristic points where the soybean plant connects to the ground. It then constructs the terrain curve between the soybeans to be harvested and the ground, as well as the line connecting points A and B, to determine the lateral rotation angle θ1 of the harvesting platform 1.

[0104] At the same time, according to the data of the second angle sensor 4, the lateral inclination angle change difference Δθ3 of the header 1 is obtained;

[0105] When |θ1-Δθ3|≤2°, the header 1 and the transverse hydraulic cylinder 5 are not adjusted;

[0106] When θ1-Δθ3>2°, the controller controls DT3 to be energized, the left end of the double-acting hydraulic cylinder is filled with oil, the piston rod is pushed out, and the left side of the cutting platform 1 is lifted, so that the cutting platform 1 matches the ground shape until |θ1-Δθ3|≤2° is satisfied;

[0107] When Δθ3-θ1>2°, the controller controls DT4 to be energized, oil enters the right end of the double-acting hydraulic cylinder, the piston rod retracts, and the left side of the cutting platform 1 falls back until |θ1-Δθ3|≤2° is satisfied, so that the cutting platform 1 matches the ground shape.

[0108] In this embodiment, when harvesting soybeans, it is necessary to avoid the header 1 from shoveling soil and missing the bottom pods, so the overall longitudinal height control target H1 of the header 1 is between 0 and h s To achieve the effect of taking both into account, set H1 = h s / 2,h s is the minimum height of the bottom pod of soybean to be harvested, so further, When harvesting rice and wheat, the stubble height needs to be controlled, and the stubble height is the control target of the overall longitudinal height of the harvesting platform 1.

[0109] The technical effects of the above technical solution are: Figure 4 As shown in the figure, in the longitudinal height control, when the difference between the actual height of the cutting platform 1 and the target height is within the allowable range, the system remains stable and does not make adjustments, which reduces unnecessary movements of the cutting platform and improves operational stability. At the same time, the precise hydraulic circuit drive system ensures that the cutting platform 1 is raised and lowered smoothly, avoiding crop shaking and mechanical shock caused by sudden height changes. As for lateral inclination control, the system adjusts the inclination of the cutting platform 1 in real time according to the terrain curve and feedback from various angle sensors, so that the cutting platform 1 maintains a stable operating posture on uneven land, enhancing the adaptability and stability of the harvester in complex terrain, and reducing the risk of uneven harvesting and mechanical failure caused by undulating terrain.

[0110] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0111] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. A harvester header height control method based on multi-source sensor information fusion is characterized in that: The following steps are involved: The harvester is integrated with a cutting platform (1), a profiling mechanism (2), a first angle sensor (3), a second angle sensor (4), a transverse hydraulic cylinder (5), a conveying trough (6), a longitudinal hydraulic cylinder (7), a third angle sensor (8) and a visual detection system; The visual inspection system collects the image information of the crops to be harvested in real time, extracts the height feature points of the crop bottom pods and the terrain feature points between the crops and the ground, constructs the terrain curve and the optimal profiling operation plane, and determines the lateral rotation angle θ1 of the header (1) and the height h of the soybean bottom pods. s ; The real-time angle change differences Δθ1 and Δθ2 on both sides of the cutting platform (1) are obtained through the profiling mechanism (2) and the first angle sensor (3), and the real-time longitudinal height H of the cutting platform (1) is calculated by combining the preset correlation coefficient and the correction coefficient; The cutting platform (1) obtains the lateral inclination angle change difference Δθ3 of the cutting platform (1) through the second angle sensor (4) to determine the lateral inclination angle of the cutting platform (1); The longitudinal height adjustment amount Δθ4 of the cutting platform (1) is obtained through the conveying trough (6) and the third angle sensor (8), and combined with the height difference between the conveying trough (6) rotating shaft and the frame and the distance from the center of the conveying trough (6) rotating shaft to the cutting blade, it is verified whether the height adjustment of the cutting platform (1) is in place; According to the determined soybean bottom pod height h s and the calculated real-time longitudinal height H of the cutting platform (1), setting a target longitudinal height H1 of the cutting platform (1), and when the difference between H and H1 exceeds a threshold, adjusting the longitudinal height of the cutting platform (1) by driving a longitudinal hydraulic cylinder (7); According to the difference between the determined lateral rotation angle θ1 of the cutting platform (1) and Δθ3 fed back by the second angle sensor (4), the lateral inclination angle of the cutting platform (1) is adjusted by driving the lateral hydraulic cylinder (5) so that the cutting platform (1) matches the terrain curve.

2. The method for controlling the height of a harvester header based on multi-source sensor information fusion according to claim 1, characterized in that: After the visual inspection system acquires the soybean plant image, it uses a deep convolutional neural network to extract the characteristic points of the soybean plant stem-pod connection and the characteristic points of the soybean plant and the ground connection. Based on the characteristic points of the soybean plant and the ground connection, points A and B are determined and a terrain curve between the soybean to be harvested and the ground is constructed. At the same time, the optimal profiling operation plane of the harvesting platform (1) is determined by connecting the two points A and B.

3. The method for controlling the height of a harvester header based on multi-source sensor information fusion according to any one of claims 1 and 2, characterized in that: Based on the characteristic points at the connection between the soybean plant stem and the bottom of the pod, determine the height difference h between each soybean plant and the line connecting points A and B. s , the lifting height of the cutting platform (1) cannot exceed h s , the angle θ1 between the line connecting points A and B and the horizontal line is the angle that the cutting platform (1) needs to rotate horizontally.

4. The method for controlling the height of a harvester header based on multi-source sensor information fusion according to claim 1, characterized in that: The calculation formula of the longitudinal real-time height H of the header (1) is: Among them, k1 is the correlation coefficient; c1 is the correction coefficient.

5. The method for controlling the height of a harvester header based on multi-source sensor information fusion according to claim 1, characterized in that: The Δθ4 obtained by the third angle sensor (8) is used to determine whether the height adjustment of the header (1) is in place, and the calculation formula is as follows: Wherein, H0 is the height difference between the rotating shaft of the conveying trough (6) and the bottom of the harvester; L0 is the distance from the center of the rotating shaft of the conveying trough (6) to the cutting knife.

6. A harvester header height control system based on multi-source sensor information fusion, applied to a harvester header height control method based on multi-source sensor information fusion as described in any one of claims 1 to 5, characterized in that: include: A harvesting platform (1) is mounted on the harvester and is used for harvesting crops, and the harvesting platform (1) is suspended on the conveying trough (6) through a hooking device; A profiling mechanism (2) is installed at the lower end of the cutting platform (1), with one set installed on each of the left and right sides of the cutting platform (1), and one side of the profiling mechanism (2) is hinged to the bottom of the cutting platform (1), and one side is suspended on the frame of the cutting platform (1); Four angle sensors are provided, namely a first angle sensor (3), a second angle sensor (4) and a third angle sensor (8), wherein the profiling mechanism (2) on both sides of the cutting platform (1) is provided with two first angle sensors (3), the cutting platform (1) is provided with a second angle sensor (4) for lateral rotation, and the conveying trough (6) is provided with a third angle sensor (8), and the angle sensors are used to measure the real-time height and lateral inclination of the cutting platform (1) and the rotation angle θ1 of the conveying trough (6); A transverse hydraulic cylinder (5) is used to swing the cutting platform (1) in the transverse direction to adjust the height; A longitudinal hydraulic cylinder (7) is hingedly mounted on the harvester frame, and drives the conveying trough (6) to rotate via the longitudinal hydraulic cylinder (7), thereby achieving the lifting and lowering of the harvesting platform (1) and changing the overall longitudinal height of the harvesting platform (1); The transverse hydraulic cylinder (5) and the longitudinal hydraulic cylinder (7) constitute a hydraulic circuit drive system, and the hydraulic circuit drive system includes two hydraulic circuits, one circuit realizes the longitudinal height control of the cutting platform (1), and the other circuit is used for the transverse height control of the cutting platform (1); The visual inspection system includes a camera (9), a graphics acquisition card and an industrial computer, and is used to collect images of crops to be harvested and analyze crop height and soybean bottom pod height information through image processing.

7. The harvester header height control system based on multi-source sensor information fusion according to claim 6, characterized in that: The hydraulic circuit drive system performs the following process: The height automatic control state of the cutting platform (1) is switched by the industrial control computer. When the longitudinal height of the cutting platform (1) is too low and the cutting platform (1) shovels the soil, DT1 is energized, the left end of the double-acting hydraulic cylinder is filled with oil, the piston rod is pushed out, and the cutting platform (1) is lifted longitudinally as a whole; When the height of the cutting platform (1) is too high and the bottom pods of soybeans are missed, the controller receives a signal to control the electromagnetic reversing valve, DT2 is energized, the right end of the double-acting hydraulic cylinder is filled with oil, the piston rod is retracted, and the cutting platform (1) is lowered longitudinally as a whole; When the harvested land is uneven with the left side higher than the right side, the controller controls DT3 to be energized, the left end of the double-acting hydraulic cylinder is filled with oil, the piston rod is pushed out, and the left side of the cutting platform (1) is lifted, so that the cutting platform (1) matches the ground shape; When the harvested land is uneven with the right side higher and the left side lower, the controller controls DT4 to be energized, oil is supplied to the right end of the double-acting hydraulic cylinder, the piston rod retracts, and the left side of the cutting platform (1) falls back, so that the cutting platform (1) matches the ground shape.

8. The harvester header height control system based on multi-source sensor information fusion according to any one of claims 6 and 7, characterized in that: The longitudinal overall height control of the header (1) is performed by executing the following process: The visual inspection system collects soybean plant images, and the industrial computer detects and calculates h in real time. s and H1; At the same time, the real-time longitudinal height H of the header (1) is calculated based on Δθ1 and Δθ2 obtained from the data of the first angle sensor (3); When |H-H1|≤10mm, the cutting platform (1) and the longitudinal hydraulic cylinder (7) are not adjusted; When H-H1>10mm, the industrial computer outputs a signal control, the solenoid valve DT2 is energized, the right end of the double-acting hydraulic cylinder is filled with oil, the piston rod is retracted, and the cutting platform (1) is lowered longitudinally as a whole until |H-H1|≤10mm is satisfied; When H1-H>10mm, the industrial computer outputs a signal control, the solenoid valve DT1 is energized, the left end of the double-acting hydraulic cylinder is filled with oil, the piston rod is pushed out, and the cutting platform (1) rises longitudinally as a whole until |H-H1|≤10mm is satisfied.

9. The harvester header height control system based on multi-source sensor information fusion according to any one of claims 6 and 7, characterized in that: The overall lateral height control of the header (1) is performed by executing the following process: After the visual inspection system acquires the soybean plant image, it uses a deep convolutional neural network to extract the characteristic points at the connection between the soybean plant and the ground, and constructs the terrain curve between the soybean to be harvested and the ground and the line connecting points A and B to determine the lateral rotation angle θ1 of the harvesting platform (1); At the same time, according to the data of the second angle sensor (4), the lateral inclination angle change difference Δθ3 of the header (1) is obtained; When |θ1-Δθ3|≤2°, the cutting platform (1) and the transverse hydraulic cylinder (5) are not adjusted; When θ1-Δθ3>2°, the controller controls DT3 to be energized, the left end of the double-acting hydraulic cylinder is supplied with oil, the piston rod is pushed out, and the left side of the cutting platform (1) is lifted, so that the cutting platform (1) matches the ground shape until |θ1-Δθ3|≤2° is satisfied; When Δθ3-θ1>2°, the controller controls DT4 to be energized, the right end of the double-acting hydraulic cylinder is supplied with oil, the piston rod is retracted, and the left side of the cutting platform (1) falls back until |θ1-Δθ3|≤2° is satisfied, so that the cutting platform (1) matches the ground shape.

10. The harvester header height control system based on multi-source sensor information fusion according to claim 6, characterized in that: The camera (9) is used to obtain images of the morphology of the soybean plants to be harvested and their height relative to the ground, the graphics acquisition card is used to collect images, the industrial control computer is used to process the images and analyze information on the height of the crops and the height of the soybean bottom pods, and the feature points of the soybean plant stem-pod connection and the feature points of the soybean plant-ground connection are extracted through a deep convolutional neural network to construct a terrain curve between the soybeans to be harvested and the ground and to determine the lateral rotation angle θ1 of the harvesting platform (1).

Images