Mountain farming machine blade guiding system with stone detection and grading protection functions and operation method of mountain farming machine blade guiding system

Abstract

The invention relates to the field of intelligent farming equipment, in particular to a mountain farming machine blade guiding system with stone detection and grading protection functions and an operation method of the mountain farming machine blade guiding system. The guiding system comprises a multi-mode sensor fusion layout module and a blade dynamic avoiding mechanism. The multi-modal sensor fusion layout module comprises a control host, an angle-adjustable cantilever support arranged above a blade of a farming machine main body, and a radar detection module; the cantilever support is provided with an image acquisition module used for scanning the terrain in front of the farming machine to generate a three-dimensional point cloud and sending the three-dimensional point cloud to the control host; the radar detection module is used for acquiring material information of obstacles in front of a running path of the farming machine and sending the material information to the control host; and the control host drives the blade dynamic avoiding mechanism to dynamically adjust the blade position of the farming machine after combined analysis according to the three-dimensional point cloud and the front obstacle substance information. And the use damage of the ploughing blade of the mountain farming machine in the practical use process is effectively reduced.

Description

TECHNICAL FIELD

[0001] The application relates to the field of intelligent farming equipment, in particular to a mountain farming machine blade guiding system with stone detection and grading protection function and a running method thereof. BACKGROUND

[0002] In complex terrains such as mountains and hills, a large number of stones and stone blocks (particle size range is usually 1cm-50cm) are generally distributed on the surface of farmland, and the density can reach several to dozens per square meter (depending on the intensity of geological activity and the degree of soil erosion). When the farming machine (such as rotary tiller, ploughing machine, etc.) is working, the blade needs to be cut into the soil frequently, and it is inevitable to collide or rub with the stones. The direct contact between the stone and the blade will cause two major problems: Mechanical damage to the blade: the hardness of the stone (Mohs hardness is usually 5-7) is much higher than that of the blade steel (HRC 40-50), and the impact or scraping will cause the blade edge to crack and the matrix to deform, and in severe cases, the blade will break and splash, directly threatening the safety of the equipment; Decrease of working efficiency and service life: the blade needs to be replaced frequently (single replacement time is about 15-30 minutes), which reduces the single day working efficiency by 30%-50%; long-term wear and tear also increases the load of the power system (the torque fluctuation amplitude increases by 20%-40%), accelerates the wear and tear of engine, transmission shaft and other components, and significantly increases the whole life cycle cost of the equipment. SUMMARY

[0003] In order to further reduce the damage of the ploughing blade of the mountain farming machine in the actual practical process, the application provides a mountain farming machine blade guiding system with stone detection and grading protection function and a running method thereof.

[0004] In the first aspect, the application provides a mountain farming machine blade guiding system with stone detection and grading protection function, which adopts the following technical scheme: A mountain farming machine blade guiding system with stone detection and grading protection function, the guiding system comprises a multi-modal sensor fusion layout module and a blade dynamic avoidance mechanism. The multi-modal sensor fusion layout module comprises a control host, a cantilever support with adjustable angle arranged above the blade of the farming machine main body and a radar detection module, an image acquisition module for scanning the terrain in front of the farming machine to generate a three-dimensional point cloud and sending it to the control host is installed on the cantilever support; the radar detection module is used to obtain the material information of the obstacle in front of the driving path of the farming machine and send it to the control host. The control host dynamically adjusts the position of the blade of the farming machine according to the three-dimensional point cloud and the material information of the obstacle in front after combined analysis.

[0005] By adopting the technical scheme, the image acquisition module can acquire the image of the farmland in front of the plowing route of the agricultural plowing machine and send the image to the control host computer, the control host computer combines the material information of the objects in the farmland obtained by the radar detection module to perform combined recognition, so that the obstacles in the farmland are intelligently recognized and avoided, and the damage of the plowing blade of the mountain agricultural plowing machine in the actual application process is reduced.

[0006] Optionally, the multi-modal sensor fusion layout module further comprises a vibration spectrum analysis sensing component. The vibration spectrum analysis sensing component comprises MEMS acceleration sensors symmetrically installed on both sides of the blade base, and a piezoelectric vibration sensor installed in the middle of the blade driving transmission shaft. The MEMS acceleration sensor is fixed to the blade connecting arm and is used for detecting high-frequency vibration caused by stone impact and sending the high-frequency vibration to the control host computer. The piezoelectric vibration sensor is used for monitoring low-frequency vibration of the blade during rotation and sending the low-frequency vibration to the control host computer, and is used for assisting the control host computer to judge the change of soil hardness.

[0007] By adopting the above technical scheme, the MEMS acceleration sensors are symmetrically installed on both sides of the blade base and are fixed to the blade connecting arm through threads, the high-frequency vibration caused by stone impact is detected, the piezoelectric vibration sensor is additionally installed in the middle of the transmission shaft, and the low-frequency vibration of the blade during rotation is monitored, which is used for assisting the judgment of the change of soil hardness.

[0008] Optionally, the blade dynamic avoidance mechanism comprises an electric lifting assembly and a blade angle fine adjustment module. The electric lifting assembly comprises an electric push rod, a base of the electric push rod is installed on the main body of the agricultural plowing machine, and an output end of the electric push rod is hingedly connected with the blade. The blade angle fine adjustment module is installed on the rotating connecting arm of the blade and is used for adjusting the cutting angle of the blade.

[0009] By adopting the above technical scheme, the blade is connected with the agricultural plowing machine rack through a hinged structure, a double-stroke electric push rod is installed below, the blade can be lifted or reset within 50 ms after receiving the control signal sent by the control host computer, and micro-terrain obstacle avoidance is realized; further, a mechanical limit switch (normally closed contact) can be integrated at the end of the stroke of the push rod to prevent overload damage (such as forced stop lifting when encountering a large stone), a harmonic reduction servo motor is additionally arranged on the blade connecting arm, the cutting angle of the blade is adjusted through a worm and gear mechanism, and small stone avoidance is realized, such as adjusting the inclination angle of the blade to bypass 5 cm of gravel.

[0010] In a second aspect, the application provides a running method of a blade guiding system of a mountain agricultural plowing machine with stone detection and grading protection function, and adopts the following technical scheme: The operation method of a mountain farming machine blade guiding system with stone detection and grading protection function comprises the following steps: multi-sensor space-time synchronous acquisition: acquiring the front terrain three-dimensional point cloud data through an image acquisition module, and synchronously outputting the stone material classification result and distance information by a radar detection module; Stone three-dimensional reconstruction and volume estimation: based on the three-dimensional point cloud data, vegetation and soil noise points are removed by a preset algorithm, and stone independent point cloud clusters are extracted; The stone point cloud cluster is subjected to Convex Hull convex hull calculation to obtain the volume of the minimum circumscribed geometric body, and the volume estimation value is corrected in combination with the material classification result; Fusion of three core parameters of stone volume (V), blade and stone relative speed (v) and soil hardness (H) to construct a risk quantification model: Wherein V th is the blade bearing limit volume threshold, v th is the safe impact speed threshold; α, β, γ are weight coefficients; According to the calculation result of the risk value R, the stone threat is divided into three levels: Level 1 (low risk): R≤0.3, stone volume is small (V<50cm 3 ) and speed is low (v<1m / s), damage probability to blade is less than 5%; Level 2 (medium risk): 0.3<R≤0.7, stone volume is medium (50cm 3 ≤V<200cm 3 ) or speed is medium (1m / s≤v<3m / s), damage probability is 30%-70%; Level 3 (high risk): R>0.7, stone volume is large (V≥200cm 3 ) or speed is high (v≥3m / s), damage probability is more than 80%; according to the hardware circuit arbitration of the risk level, the host outputs the corresponding control command to the blade dynamic avoidance mechanism.

[0011] By adopting the above technical scheme, the method realizes the technical leap from passive protection to active adaptation through the "perception-decision-execution-learning" closed-loop architecture, can significantly reduce the blade damage rate, and can improve the operation efficiency of the mountain hilly farming machine by 20%-30%. On the one hand, the risk value is calculated by coupling the stone volume, relative speed and soil hardness, which further improves the accuracy compared with single parameter judgment; on the other hand, the risk model weight is updated online and iteratively to adapt to different farmland environments (such as mountain / sand / clay alternating plots), which breaks through the limitations of the fixed threshold scheme.

[0012] Optionally, the step of arbitrating by the hardware circuit according to the risk level and controlling the host computer to output a corresponding control instruction to the blade dynamic avoidance mechanism specifically comprises the following sub-steps: Level 1: trigger blade angle fine adjustment, record stone position for path optimization; Level 2: start the electric push rod to lift the blade, and reduce the tillage speed at the same time; Level 3: immediately stop the blade rotation, and trigger the agricultural cultivator ECU to execute emergency braking.

[0013] By adopting the above technical solution, different levels of protection functions can be implemented on the blade according to different actual tillage conditions, the use loss of the blade can be reduced without affecting the tillage efficiency, and thus the use time of the blade is prolonged.

[0014] Optionally, the method further comprises the following steps: Short-term path re-planning: mapping the detected stone position to the digital farmland map, generating a local obstacle avoidance path by a preset algorithm, and updating to the agricultural cultivator navigation system; Long-term adaptive learning: Constructing a stone distribution-soil hardness correlation database to record the environmental parameters of each detection event; Using a Q-learning reinforcement learning algorithm to dynamically optimize the risk quantification model weight coefficients (α, β, γ), and the objective function is to minimize the blade damage probability and tillage efficiency loss.

[0015] Optionally, the step of short-term path re-planning, mapping the detected stone position to the digital farmland map, generating a local obstacle avoidance path by a preset algorithm, and updating to the agricultural cultivator navigation system specifically comprises the following sub-steps: Converting the detected stone point cloud cluster center coordinates to the agricultural cultivator body coordinate system and the global map coordinate system in turn, and finally mapping to the grid index of the digital farmland map; Dynamic obstacle map updating and path search space construction; In the local map window, based on the dynamic obstacle mask and the multi-objective cost function, an optimal obstacle avoidance path is searched based on a preset algorithm; The discrete path point sequence output by the preset algorithm is smoothed, and a speed planning instruction is generated.

[0016] By adopting the above technical solution, for the obstacles (such as stones, gravel, wooden piles, etc.) encountered in the tillage process, a tillage route that can directly avoid the obstacles is planned through digital map imaging, so as to further improve the tillage efficiency.

[0017] In a third aspect, the computer device provided by the present application adopts the following technical solution: The computer device comprises a processor, a memory and a computer program stored in the memory and executable on the processor, and the processor implements the operation method of the blade guiding system of the mountain farming machine with the stone block detection and grading protection function when executing the computer program.

[0018] By adopting the technical scheme, the computer device capable of executing the operation method of the blade guiding system of the mountain farming machine with the stone block detection and grading protection function is provided.

[0019] In a fourth aspect, the computer readable storage medium provided by the present application adopts the following technical scheme: The computer readable storage medium stores a computer program; and the computer program is executed by a processor to implement the operation method of the blade guiding system of the mountain farming machine with the stone block detection and grading protection function.

[0020] By adopting the technical scheme, the carrier of the computer program of the operation method of the blade guiding system of the mountain farming machine with the stone block detection and grading protection function is provided.

[0021] In summary, the present application includes the following at least beneficial technical effects: 1. Based on the image acquisition module, the farmland image in front of the farming machine plowing route can be acquired and sent to the control host, the control host combines the material information of each object in the farmland obtained by the radar detection module to perform combined recognition, so as to intelligently identify / avoid the obstacles in the farmland; thereby reducing the use damage of the farmland blade of the mountain farming machine in the actual practical process; 2. The method realizes the technical leap from passive protection to active adaptation through the "perception-decision-execution-learning" closed loop architecture, can significantly reduce the blade damage rate, and improve the operation efficiency of the mountain and hilly farming machine by 20%-30%. On the one hand, the risk value is calculated by coupling the volume, relative speed and soil hardness of the stone, which further improves the accuracy compared with single parameter judgment; on the other hand, the risk model weight is updated online and iteratively, which adapts to different farmland environments (such as mountain / sand / clay alternating plots), and breaks through the limitations of the fixed threshold scheme. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 is a short-term path re-planning flowchart of the farming machine based on multi-modal perception of the present application; Figure 2 is a flowchart of step S400 of the present application.

[0023] REFERENCE SIGNS: 1. Control host; 2. Radar detection module; 3. Image acquisition module. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application and not to limit the present application. Figures 1-2 The present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application and not to limit the present application.

[0025] The present application discloses a mountain farming machine blade guiding system with stone detection and hierarchical protection function.

[0026] With reference to Figure 1 and Figure 2 , a mountain farming machine blade guiding system with stone detection and hierarchical protection function, comprising a multi-modal sensor fusion layout module and a blade dynamic avoidance mechanism installed on the body of the farming machine; In the present application, the multi-modal sensor fusion layout module comprises a control host 1, an adjustable-angle cantilever bracket arranged above the blade of the main body of the farming machine, and a radar detection module 2. An image acquisition module 3 for scanning the terrain in front of the farming machine to generate a three-dimensional point cloud and sending it to the control host 1 is installed on the cantilever bracket. The radar detection module 2 is used to obtain the material information of the obstacles in front of the driving path of the farming machine and send it to the control host 1. For example, an adjustable-angle aluminum cantilever bracket can be fixedly installed in front of the blade of the farming machine (10-20 cm away from the cutting surface of the blade), and the image acquisition module 3 and the radar detection module 2 are synchronously installed on the top of the cantilever bracket. In the present embodiment, the image acquisition module 3 can be a fixed solid-state LiDAR sensor (horizontal field of view ≥ 270°, vertical field of view ≥ 90°, resolution ≤ 5mm@10m), which can scan the terrain in front to generate a three-dimensional point cloud.

[0027] In the present example, the radar detection module 2 uses FMCW millimeter wave radar (detection distance 0.5-30m, angle resolution ≤ 1°), which can penetrate vegetation barriers to detect stone materials (distinguish between rocks / metal), and is spatiotemporally synchronized with LiDAR data (through sharing the timestamp of the IMU inertial measurement unit).

[0028] In summary, the control host 1 drives the blade dynamic avoidance mechanism to dynamically adjust the position of the blade of the farming machine based on the three-dimensional point cloud and the material information of the obstacles in front. Based on the image acquisition module 3, the image of the farmland in front of the farming route can be acquired and sent to the control host 1, and the control host 1 combines the material information of each object in the farmland obtained by the radar detection module 2 to identify and avoid the obstacles in the farmland intelligently; thereby reducing the damage of the farmland blade of the mountain farming machine in the actual practical process.

[0029] Optionally, the multi-modal sensor fusion layout module further comprises a vibration spectrum analysis sensor assembly; in this embodiment, the vibration spectrum analysis sensor assembly comprises MEMS acceleration sensors symmetrically installed on both sides of the blade base, and a piezoelectric vibration sensor installed in the middle of the blade drive transmission shaft; The MEMS acceleration sensors are fixed to the blade connecting arms and used to detect high-frequency vibrations caused by stone impacts and send them to the control host 1; the piezoelectric vibration sensor is used to monitor low-frequency vibrations when the blade is rotating and send them to the control host 1, which is used to assist the control host 1 in judging changes in soil hardness.

[0030] In summary, the MEMS acceleration sensors are symmetrically installed on both sides of the blade base and fixed to the blade connecting arms by threads, which are used to detect high-frequency vibrations caused by stone impacts; the piezoelectric vibration sensor is installed in the middle of the transmission shaft, which is used to monitor low-frequency vibrations when the blade is rotating and assist in judging changes in soil hardness.

[0031] With reference to Figure 1 and Figure 2 , the blade dynamic avoidance mechanism comprises an electric lifting assembly and a blade angle fine-tuning module. The electric lifting assembly comprises an electric push rod, the base of which is installed on the main body of the farming machine, and the output end of which is hinged to the blade; the blade angle fine-tuning module is installed on the rotating connecting arm of the blade and used to adjust the cutting angle of the blade.

[0032] The blade is connected to the farming machine frame through a hinged structure, and a double-stroke electric push rod is installed below, which can lift or reset the blade within 50 ms after receiving the control signal sent by the control host 1, thereby realizing micro-terrain obstacle avoidance; further, a mechanical limit switch (normally closed contact) can be integrated at the end of the push rod stroke to prevent overload damage (such as forced stop lifting when encountering large stones); a harmonic reduction servo motor is additionally provided on the blade connecting arm, which adjusts the cutting angle of the blade through a worm and gear mechanism, and is used for small stone avoidance, such as fine-tuning the blade inclination angle to bypass 5 cm of gravel.

[0033] Based on the same design concept, the embodiment also discloses a running method of a mountain farming machine blade guiding system with stone detection and grading protection function.

[0034] With reference to Figure 1 and 2 , the running method of the mountain farming machine blade guiding system with stone detection and grading protection function comprises the following steps: S100: multi-sensor spatiotemporal synchronous acquisition: the image acquisition module 3 acquires the front terrain three-dimensional point cloud data, and the radar detection module 2 synchronously outputs the stone material classification result and distance information.

[0035] Specifically, the three-dimensional point cloud data, stone material information and distance information are all sent to the control host 1, the control host 1 combines the material information of each object in the cultivated land obtained by the radar detection module 2 to perform combined recognition, so as to intelligently identify / avoid the obstacles in the cultivated land; thereby reducing the use damage of the cultivated land blade of the mountain farmland machine in the actual practical process.

[0036] S200: Stone three-dimensional reconstruction and volume estimation: based on three-dimensional point cloud data, vegetation and soil noise points are removed by a preset algorithm, and stone independent point cloud clusters are extracted; Specifically, when it is detected that the stone needs to be locally avoided, the system generates a safe obstacle avoidance path within 100ms through multi-source data fusion and real-time path optimization and issues the path to the farmland machine navigation system, and the specific steps are as follows: S210: Stone position coordinate conversion and map matching.

[0037] For example, the stone point cloud cluster center coordinates detected by the LiDAR are converted to the farmland machine body coordinate system and the global map coordinate system in turn, and finally mapped to the grid index of the farmland digital map, and the specific process is as follows: S211: LiDAR point cloud coordinate conversion (local→machine body coordinate system); Specifically, the stone point cloud cluster center coordinates detected by the LiDAR (LiDAR local coordinate system, with the origin at the sensor installation base) are converted to the farmland machine body coordinate system (with the origin at the center of gravity of the agricultural machine) through the hand-eye calibration matrix, so as to eliminate the sensor installation offset error (calibration accuracy ±1cm).

[0038] Let the LiDAR coordinate system be L, the machine body coordinate system be B, and the conversion matrix be T LB (4×4 homogeneous transformation matrix), then the coordinates of the stone in the machine body coordinate system are: P B =T LB ·P L S212: LiDAR point cloud coordinate conversion machine body coordinate system→global map coordinate system (machine body→global); Specifically, the global coordinates (longitude λ, latitude φ, and height h, accuracy ±2cm) of the current position of the farmland machine are obtained by the RTK-GNSS receiver, and the IMU attitude angle is combined to construct the rotation translation matrix T BG from the machine body coordinate system to the global map coordinate system G, and finally the coordinates of the stone in the global map are obtained: P G =T BG ·P B Further, Extended Kalman Filter (EKF) is used to fuse RTK-GNSS and IMU data to compensate the positioning drift caused by the bumps of agricultural machinery (error < 3cm).

[0039] S213: Global coordinates to digital map grid index (global to grid); Specifically, the stone global coordinates P G Convert to the grid index (x, y) of the farmland digital map (map resolution 0.05m / pixel), determine the grid position of the stone center point through bilinear interpolation, and mark it as an obstacle grid (attribute = high-hardness obstacle).

[0040] S220: Dynamic obstacle map update and path search space construction.

[0041] Update the local map window based on the stone position, and dynamically generate the passable area mask, with the specific process as follows: S221: Local map window extraction.

[0042] Specifically, a 10m x 10m local map window (covering the typical plowing path length) is intercepted with the current position of the agricultural machine as the center, containing the stone obstacle and its surrounding environment (such as the plowed area and the boundary of the field ridge). Further, a sliding window mechanism can also be used, with the window center updated in real time as the agricultural machine moves (update frequency 10Hz), avoiding repeated calculation of the global map.

[0043] S222: Dynamic obstacle inflation processing.

[0044] Perform Euclidean Distance Transformation (EDT) on the stone obstacle grid, and calculate the safe inflation radius r according to the stone volume V: Where k = 0.2 (proportionality coefficient, empirical value), c = 0.1m (basic safety margin). After inflation, the obstacle grid range covers the stone projection area and the surrounding loose soil (to avoid blade edge scratching).

[0045] S223: Passable area mask generation.

[0046] Combine the pre-defined attributes of the farmland digital map (such as field ridge and ditch, which are not passable), merge the inflated obstacle grid with the static non-passable area, and generate a dynamic passable mask map (BinaryMask), marking all feasible grids (value 1) and prohibited grids (value 0).

[0047] S300: Perform Convex Hull convex hull calculation on the stone point cloud cluster to obtain its minimum circumscribed geometric volume, and correct the volume estimation value combined with the material classification result; The three core parameters of fused stone block volume (V), blade and stone block relative speed (v), and soil hardness (H) are used to construct a risk quantification model: wherein V th is the blade bearing limit volume threshold, v th is the safe impact speed threshold; and α, β, and γ are weight coefficients. According to the calculation result of the risk value R, the stone block threat is divided into three levels: Level 1 (low risk): R≤0.3, the stone block volume is small (V<50cm 3 ) and the speed is low (v<1m / s), the damage probability to the blade is less than 5%; Level 2 (medium risk): 0.3<R≤0.7, the stone block volume is medium (50cm 3 ≤V<200cm 3 ) or the speed is medium (1m / s≤v<3m / s), the damage probability is 30% to 70%; Level 3 (high risk): R>0.7, the stone block volume is large (V≥200cm 3 ) or the speed is high (v≥3m / s), the damage probability is more than 80%; S400: according to the hardware circuit arbitration of the risk level, the control host outputs corresponding control instructions to the blade dynamic avoidance mechanism.

[0048] Level 1: trigger the blade angle fine adjustment, and record the stone block position for path optimization; Specifically, the step S400 includes the following sub-steps: S410: short-term path re-planning; map the detected stone block position to the digital farmland map, generate a local obstacle avoidance path through a preset algorithm, and update to the agricultural machine navigation system; S420: sequentially convert the center coordinates of the detected stone block point cloud cluster to the agricultural machine body coordinate system and the global map coordinate system, and finally map to the grid index of the digital farmland map; S430: dynamic obstacle map update and path search space construction; S440: in the local map window, based on the dynamic obstacle mask and the multi-target cost function, search for the optimal obstacle avoidance path based on the preset algorithm; Specifically, the preset algorithm can be an A* search algorithm, a time elastic band (TEB) algorithm, or other intelligent algorithms in the same field that can plan paths, all of which can be adaptively adjusted and used.

[0049] S450: smooth the discrete path point sequence output by the preset algorithm, and generate a speed planning instruction.

[0050] For example, the output discrete path point sequence is fitted with a cubic B-spline curve to generate a continuous smooth trajectory (curvature is continuous, and the maximum curvature radius is greater than or equal to 0.5 m) to avoid impact load caused by path mutation of the blade. The smoothed path point sequence (including speed instructions) is transmitted to the agricultural machine ECU through the CANFD bus (period 10 ms) and the global path planning of the navigation system is updated (the local path covers the corresponding section of the original path).

[0051] S460: Long-term adaptive learning: S470: Construct a stone distribution-soil hardness correlation database to record the environmental parameters of each detection event. S480: Dynamically optimize the risk quantification model weight coefficients (a, b, g) using the Q-learning reinforcement learning algorithm, and the objective function is to minimize the blade damage probability and the loss of farming efficiency.

[0052] Level2: Start the electric push rod to lift the blade, and at the same time, reduce the farming speed; Level3: Immediately stop the blade rotation, and trigger the agricultural machine ECU to execute emergency braking.

[0053] The application also provides a computer readable storage medium, which stores a computer program capable of being loaded and executed by a processor to implement the above steps.

[0054] The computer readable storage medium includes, for example, a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0055] Based on the same inventive concept, the embodiments of the application provide a computer device, which includes a memory and a processor, and the memory stores a computer program capable of being loaded and executed by the processor to implement the above method.

[0056] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional modules is taken as an example for illustration, and in actual application, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the above-described system, device and unit can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0057] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other manners. For example, the described device embodiment is merely illustrative. For example, the division of the modules or units is merely logical function division, and there can be another division manner in actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections can be indirect couplings or communication connections through some interfaces, devices or units, and can be in electrical, mechanical or other forms.

[0058] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, i.e., can be located in one place, or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0059] In addition, each functional unit in the various embodiments of the present application can be integrated into one processing unit, or each unit can exist physically, or two or more units can be integrated into one unit. The integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0060] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the part that makes a contribution to the prior art, or all or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to execute all or part of the steps of the methods described in the various embodiments of the present application. The foregoing storage medium includes: U disk, mobile hard disk, read-only memory, random access memory, magnetic disk or optical disk, and various media that can store program codes.

[0061] The above-described embodiments are merely used to describe the technical solutions of the present application in detail, but the description of the above embodiments is only used to help understand the method and core idea of the present application, and should not be construed as limiting the present application. Those skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application.

Claims

1. A blade guiding system for a mountain agricultural machine with stone detection and graded protection functions, characterized in that: The guidance system includes a multimodal sensor fusion layout module and a blade dynamic avoidance mechanism; The multimodal sensor fusion layout module includes a control host (1), an adjustable cantilever bracket set above the blades of the main body of the agricultural machine, and a radar detection module (2). The cantilever bracket is equipped with an image acquisition module (3) for scanning the terrain in front of the agricultural machine to generate a three-dimensional point cloud and sending it to the control host (1). The radar detection module (2) is used to acquire the material information of obstacles in front of the agricultural machine's driving path and send it to the control host (1). The control host (1) analyzes the three-dimensional point cloud and the information of obstacles in front and then drives the blade dynamic avoidance mechanism to dynamically adjust the position of the blade of the agricultural tiller.

2. The blade guiding system for mountain agricultural machinery with stone detection and graded protection functions according to claim 1, characterized in that, The multimodal sensor fusion layout module also includes a vibration spectrum analysis sensing component; The vibration spectrum analysis sensing component includes a MEMS accelerometer symmetrically mounted on both sides of the blade base and a piezoelectric vibration sensor mounted in the middle of the blade drive shaft. The MEMS accelerometer is fixed to the blade connecting arm and is used to detect the high-frequency vibration generated by the impact of the stone and send it to the control host (1); The piezoelectric vibration sensor is used to monitor the low-frequency vibration when the blade rotates and send it to the control host (1) to assist the control host (1) in judging the change in soil hardness.

3. The blade guiding system for mountain agricultural machinery with stone detection and graded protection functions according to claim 1, characterized in that, The blade dynamic avoidance mechanism includes an electric lifting assembly and a blade angle fine-tuning module; The electric lifting assembly includes an electric push rod, the base of which is mounted on the main body of the agricultural tiller, and the output end of which is hinged to the blade. The blade angle fine-tuning module is installed on the rotating connecting arm of the blade and is used to adjust the blade cutting angle.

4. A method for operating a blade guiding system for a mountain agricultural machine with stone detection and graded protection functions, characterized in that, Includes the following steps: Multi-sensor spatiotemporal synchronous acquisition: The image acquisition module acquires three-dimensional point cloud data of the terrain ahead, and the radar detection module simultaneously outputs the stone material classification results and distance information; 3D reconstruction and volume estimation of stones: Based on 3D point cloud data, vegetation and soil noise points are removed by a preset algorithm to extract independent point cloud clusters of stones; Convex Hull calculation is performed on the point cloud cluster of rocks to obtain its minimum bounding geometry volume, and the volume estimate is corrected by combining the material classification results. By integrating three core parameters—stone volume (V), relative velocity between the blade and the stone (v), and soil hardness (H)—a risk quantification model is constructed. Where V th For the blade's load-bearing limit volume threshold, v th α, β, and γ are the safe impact velocity thresholds; α, β, and γ are weighting coefficients. Based on the risk value R, the rock threat is divided into three levels: Level 1 (Low Risk): R≤0.3, small stone volume (V<50cm) 3 Furthermore, the speed is low (v<1m / s), resulting in a damage probability of <5% to the blade; Level 2 (Medium Risk): 0.3 < R ≤ 0.7, the volume of the rock is medium (50 cm 3 ≤ V < 200 cm 3 ) or the speed is medium (1 m / s ≤ v < 3 m / s), and the injury probability is 30% - 70%; Level 3 (High Risk): R > 0.7, large stone volume (V ≥ 200cm) 3 If the speed is high (v≥3m / s), the probability of damage is >80%; based on the hardware circuit arbitration of the risk level, the control host outputs the corresponding control command to the blade dynamic avoidance mechanism.

5. The operating method of the blade guiding system for a mountain agricultural machine with stone detection and graded protection functions according to claim 4, characterized in that, The step of arbitrating the hardware circuit according to the risk level and outputting corresponding control commands from the control host to the blade dynamic avoidance mechanism specifically includes the following sub-steps: Level 1: Triggers fine-tuning of blade angle and records stone positions for path optimization; Level 2: Activate the electric push rod to raise the blades while reducing the tillage speed; Level 3: Immediately stop blade rotation and trigger the tiller ECU to perform emergency braking.

6. The operating method of the blade guiding system for a mountain agricultural machine with stone detection and graded protection functions according to claim 5, characterized in that, It also includes the following steps: Short-term path replanning; the detected stone locations are mapped to a digital farmland map, and a local obstacle avoidance path is generated using a preset algorithm and updated to the agricultural machinery navigation system; Long-term adaptive learning: Construct a database linking stone distribution and soil hardness, and record environmental parameters for each detection event; The Q-learning reinforcement learning algorithm is used to dynamically optimize the weight coefficients (α,β,γ) of the risk quantification model. The objective function is to minimize the probability of blade damage and the loss of tillage efficiency.

7. The operating method of the blade guiding system for a mountain agricultural machine with stone detection and graded protection functions according to claim 1, characterized in that: The short-term path replanning, which maps the detected stone locations to a digital farmland map, generates a local obstacle avoidance path using a preset algorithm, and updates the agricultural machinery navigation system, specifically includes the following sub-steps: The center coordinates of the detected stone point cloud clusters are sequentially transformed to the coordinate system of the agricultural machine body and the global map coordinate system, and finally mapped to the raster index of the farmland digital map. Dynamic obstacle map updating and path search space construction; Within a local map window, the optimal obstacle avoidance path is searched based on a preset algorithm using dynamic obstacle masks and multi-objective cost functions. The discrete path point sequence output by the preset algorithm is smoothed and a speed planning instruction is generated.

8. A computer device, characterized in that, The system includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the operation method of the mountain farm blade guide system with stone detection and graded protection functions as described in any one of claims 4-7.

9. A computer-readable storage medium, characterized in that, The computer storage medium stores a computer program; when the computer program is executed by the processor, it implements the operation method of the mountain agricultural machine blade guide system with stone detection and graded protection function as described in any one of claims 4-7.

Images