Hilly mountain land weeding robot and working method

By integrating infrared sensors, binocular cameras, and oscillation sensors, along with lifting components, the problem of weeding in orchards with varying weed heights and complex terrain has been solved, achieving efficient and stable weeding results.

CN119032729BActive Publication Date: 2025-11-04SHANDONG ACADEMY OF AGRICULTURAL MACHINERY SCIENCES
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202411359407.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-27
Publication Date
2025-11-04
Estimated Expiration
2044-09-27

AI Technical Summary

Technical Problem

Existing small weeding machines are unable to cope with the differences in weed height and complex terrain in orchards, resulting in a large amount of manpower still being required after weeding operations.

Method used

Integrating infrared sensors, binocular cameras, and vibration sensors, combined with lifting components, it achieves precise perception and dynamic response to weed height and terrain, enabling intelligent weed removal via a lawnmower.

Benefits of technology

It achieves efficient removal of weeds in the orchard, ensures the stability and precision of the weeding process, reduces manpower input, and prevents weeds from regrowing.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119032729B_ABST
    Figure CN119032729B_ABST
Patent Text Reader

Abstract

A hilly mountainous area weeding robot and working method, for adjusting the blade height according to the size, height and ground concave-convex situation of weeds, solves the problem that the existing weeding machinery cannot clean low weeds. The weeding robot comprises a robot shell, the front end of the robot shell is provided with a steering wheel, the rear end of the robot shell is provided with a driving wheel, the front side wall of the robot shell is provided with an infrared sensor for detecting obstacles and a binocular camera for shooting ground images, the lower portion of the front end of the robot shell is provided with a lifting plate, the lifting plate and the robot shell are provided with a lifting assembly for adjusting the height of the lifting plate, the lifting assembly adjusts the height of the lifting plate according to the height of weeds, and a rotary weeding cutter is rotatably installed on the lifting plate through a cutter shaft to realize rotary weeding. The present application can adjust the cutter height according to the height of weeds to adapt to the weeding of weeds of different heights.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of agricultural machinery technology, specifically to a weeding robot and its working method for hilly and mountainous areas. Background Technology

[0002] Fruit trees generally have large canopies and small trunks, making them unsuitable for large-scale weeding machinery, while manual weeding is time-consuming and labor-intensive. A smart weeding robot, disclosed in CN112567960A, includes a vehicle body with wheels at the bottom and a camera at the top. A second drive weeding motor is located at the bottom of the vehicle body, and blades are connected to the second drive weeding motor, which rotates the blades to perform the weeding action. A grass-gathering plate is located at the front of the vehicle body, gradually opening outwards from the front to gather target grass inwards. A control system is connected to the blades, camera, and wheels to control the working status of the blades and camera, as well as the vehicle's travel path. The first drive weeding motor drives the wheels and plans the robot's path based on feedback from the control system. The aforementioned robots rely on mowers at a fixed height for operation, making it difficult to cope with the varying heights of weeds and complex terrain in orchards. Furthermore, they lack effective methods for identifying and treating weeds, resulting in a significant need for manpower for field management after weeding. Summary of the Invention

[0003] The purpose of this invention is to provide a hilly and mountainous weeding robot and its working method, which adjusts the blade height according to the size and height of the weeds and the unevenness of the ground, solving the problem that existing small weeding machines cannot clean low-lying weeds; by integrating infrared sensors, binocular cameras, vibration sensors and intelligent lifting components, it can achieve accurate perception and dynamic response to the size and height of weeds and complex terrain in orchards.

[0004] The technical solution adopted by this invention to solve its technical problem is as follows: a hilly and mountainous weeding robot, including a robot shell, a steering wheel at the front end of the robot shell, an oscillation sensor for detecting oscillation signals on the steering wheel and drive wheel, a drive wheel at the rear end of the robot shell, an infrared sensor for detecting obstacles on the front side wall of the robot shell, and a binocular camera for capturing ground images. The binocular camera is used to distinguish and separate weed areas in the image, and then perform in-depth analysis on the extracted weed area images, performing multiple weed feature extractions to construct a detailed weed feature set; a lifting plate is located below the front end of the robot shell, and a lifting assembly for adjusting the height of the lifting plate is located between the lifting plate and the robot shell. The lifting assembly adjusts the height of the lifting plate according to the height of the weeds. A weed cutter that performs weeding in a rotating manner is rotatably mounted on the lifting plate via a weed cutter shaft; the oscillation sensor collects the oscillation signal between the robot and the terrain in real time, and accurately identifies complex terrain by combining the oscillation signal with the image data collected by the binocular camera, and makes timely dynamic response adjustments when the terrain changes, ensuring stability and operational accuracy during the weeding process.

[0005] Furthermore, the lifting assembly includes two symmetrically arranged telescopic cylinders, the cylinder barrels of which are fixedly connected to the robot shell, and the piston rods of which are fixedly connected to the lifting plate.

[0006] Furthermore, the lawnmower consists of two symmetrically arranged on the lifting plate.

[0007] Furthermore, a weeding motor is fixed on the lifting plate, and a driving bevel gear is fixed on the output shaft of the weeding motor. The mower and the driven bevel gear are coaxially arranged, and the driven bevel gear meshes with the driving bevel gear. A protective cover is fixed on the lifting plate to cover the weeding motor, the driving bevel gear and the driven bevel gear.

[0008] Furthermore, the robot shell has a negative pressure pump in the middle and a storage compartment at the rear end, and the negative pressure pump is connected to the storage compartment through a pipe.

[0009] Furthermore, the storage chamber contains a crushing assembly, which includes two crushing rollers rotating in opposite directions. The outer wall of the crushing rollers has crushing teeth, and the crushing rollers are rotatably installed inside the storage chamber. The bottom of the robot shell has a discharge hole that communicates with the storage chamber for discharging the crushed weeds.

[0010] Furthermore, a drive gear is fixed on the roller shaft of one of the crushing rollers, and a driven gear is fixed on the roller shaft of the other crushing roller. The drive gear and the driven gear mesh and are both located outside the robot shell. The drive gear rotates synchronously with the drive wheel.

[0011] Furthermore, the drive gear is coaxially arranged with the friction roller, the friction roller is in contact and pressed with the drive wheel, and the drive wheel rotates to drive the friction roller to rotate.

[0012] Furthermore, the lawnmower includes a blade disc and blades fixed to the outer wall of the blade disc. The blades are arc-shaped blades with their protruding portions facing downwards. The lower end of the lawnmower shaft extends into the blade disc. The lower end of the lawnmower shaft has a sliding plate. The sliding plate is slidably connected to a groove in the blade disc and slides in the up-down direction. A spring is located between the bottom of the sliding plate and the groove. The blade height is lowest when the spring is at its natural length.

[0013] The present invention also provides a method for operating a weeding robot in hilly and mountainous areas, comprising the following steps:

[0014] S1. The robot autonomously navigates along a set route in the orchard and uses the infrared sensor to dynamically scan and measure the time difference and phase difference of the transmitted and received signals to monitor the specific location and distance of obstacles in real time.

[0015] S2. High-precision acquisition of orchard images is achieved through the binocular camera. The left and right images are registered and processed using a stereo matching algorithm to generate a precise depth map of the orchard scene. Based on image preprocessing and target detection technology, the extracted weed area images are analyzed in depth to extract multi-dimensional weed features and construct a complete weed feature set. When the detected ground weed area exceeds a specified threshold, the height of the lifting plate is intelligently adjusted according to the depth map information provided by the binocular camera, and the lawnmower is started to rotate to clear the ground weeds.

[0016] S3. The robot uses a multi-input, multi-output vibration model for terrain classification. Through the oscillation sensor and binocular camera mounted on the robot, it collects oscillation signals and image data in real time. The oscillation signals are used to supplement and verify the terrain information in the image data. When the robot travels in complex terrain, it adjusts the navigation strategy and control parameters according to the fusion result of image data and oscillation signals to achieve accurate terrain classification and stable navigation.

[0017] S4. Start the negative pressure pump to suck the cut weeds into the storage bin, crush the weeds through the crushing component, and discharge the crushed weeds out of the storage bin through the discharge hole.

[0018] The beneficial effects of this invention are as follows: By integrating multiple technologies such as infrared sensors, binocular cameras, and oscillation sensors, this invention enables robots to intelligently avoid obstacles and efficiently remove weeds in orchard environments; the infrared sensors accurately monitor the location of obstacles, and the binocular cameras, combined with stereo matching algorithms, obtain orchard depth maps, identify weeds, and intelligently adjust the lifting components according to their height to ensure efficient removal of weeds at different heights; the fusion of oscillation signals and image data allows the robot to adjust its navigation strategy in real time, achieving accurate terrain classification; the lawnmower removes weeds from the ground by rotating, enabling rapid weed removal; the negative pressure pump adsorbs weeds, and the weeds are crushed in the storage bin by a crushing component, preventing weed regrowth. Attached Figure Description

[0019] Figure 1 This is one of the three-dimensional diagrams of the present invention;

[0020] Figure 2 This is the second three-dimensional diagram of the present invention;

[0021] Figure 3 A 3D diagram of the internal structure of the protective shield;

[0022] Figure 4 for Figure 3 Enlarged view of section A in the image;

[0023] Figure 5 This is a three-dimensional view of the internal structure of the outer shell of the present invention;

[0024] Figure 6 A 3D view of the broken components;

[0025] Figure 7 A cross-sectional view of another assembly method of the lawnmower and the lawnmower shaft;

[0026] Figure 8 This is a schematic diagram of the local correlation algorithm for binocular cameras.

[0027] Figure 9 This is a schematic diagram of the three-degree-of-freedom vibration of the robot according to the present invention;

[0028] In the diagram: 1. Robot shell, 2. Steering wheel, 3. Drive wheel, 4. Infrared sensor, 5. Binocular camera, 6. Lifting plate, 7. Lawn mower, 71. Cutting disc, 72. Blade, 73. Slide, 8. Telescopic cylinder, 9. Protective cover, 10. Driven bevel gear, 11. Driven bevel gear, 12. Negative pressure pump, 13. Storage bin, 14. Discharge port, 15. Crushing roller, 16. Driven gear, 17. Driven gear, 18. Friction roller, 19. Weeding motor, 20. Output shaft, 21. Lawn mower shaft, 22. Pipe, 23. Crushing teeth, 24. Slide plate, 25. Spring. Detailed Implementation

[0029] like Figures 1 to 9 As shown, the orchard obstacle avoidance and weeding robot of the present invention includes a robot shell 1, steering wheels 2, drive wheels 3, infrared sensors 4, binocular cameras 5, lifting plates 6, lawnmowers 7, lifting components, protective covers 9, drive mechanisms, vibration sensors, speed sensors, crushing components, and adsorption components. The present invention will be described in detail below with reference to the accompanying drawings.

[0030] like Figures 1 to 7 As shown, a hilly weeding robot includes a robot shell 1. The front end of the robot shell 1 has a steering wheel 2, and the rear end has a drive wheel 3. The drive wheel 3 is driven to rotate by a walking motor mounted on the robot shell 1, thereby driving the movement of the entire robot. The robot shell 1 has a steering mechanism for adjusting the angle of the steering wheel 2; the robot's steering is achieved by changing the angle of the steering wheel 2 (i.e., the steering wheel 2 deflects). The drive wheel 3, the driving method of the drive wheel 3, the steering wheel 2, and the deflection method of the steering wheel 2 are all existing technologies and will not be described in detail. The steering wheel 2 and the drive wheel 3 have oscillation sensors, and the robot shell 1 is equipped with a speed sensor.

[0031] like Figure 1 As shown, the front sidewall of the robot's outer shell 1 has an infrared sensor 4 for obstacle detection. The infrared sensor 4 performs dynamic scanning, and based on the measurement of the time difference and phase difference between the transmitted and received signals, it calculates the specific position and distance of obstacles in real time, accurately detecting the spatial layout of the orchard environment. The data acquired by the infrared sensor 4 is processed by the controller, which dynamically updates the robot's travel path based on the obstacle position information, thereby avoiding obstacles and ensuring the continuity of the navigation process.

[0032] The front sidewall of the robot's outer shell 1 is equipped with a binocular camera 5 for capturing ground images and determining the presence and height of weeds. The binocular camera 5 enables high-precision acquisition of orchard images. A stereo matching algorithm is used to register and process the left and right images, generating a precise depth map of the orchard scene. Based on image preprocessing and target detection technology, the robot can perform in-depth analysis of the extracted weed area images, extracting multi-dimensional weed features and constructing a complete weed feature set.

[0033] like Figure 9As shown, the robot of this invention includes a multi-input, multi-output three-degree-of-freedom vibration model, which contains four input signals and three output signals. During robot movement, the controller extracts the four input signals from the oscillation sensors connected to the robot's steering wheels and drive wheels. The output signals include the pitch rate, roll rate, and vertical rate captured by the robot's speed sensors, used to quantify the robot's dynamic behavior and response during movement, describing the vehicle's rotational speed around the horizontal and vertical axes, and the change in acceleration in the vertical direction. Specific implementation details are as follows:

[0034] (1) High-precision image acquisition by binocular camera

[0035] The robot's front-end binocular camera 5 simultaneously captures images of the orchard scene from both left and right perspectives, obtaining two images with parallax. An oscillation sensor records the oscillation signals as the robot contacts the ground, recording the unevenness of the ground and the robot's motion state.

[0036] (2) Image registration using stereo matching algorithm

[0037] By using SIFT image registration technology based on feature points to process two images, the same objects in the left and right views are placed in the correct corresponding positions, reducing the offset caused by camera differences or motion.

[0038] (3) Depth map generation

[0039] 1) Matching cost calculation

[0040] like Figure 8 As shown, the pixel gradient difference method is used to calculate the matching cost of each pixel in the left and right images and determine which pixel in the two images corresponds to the same object.

[0041] 2) Matching cost aggregation

[0042] By aggregating the cost information of neighboring pixels and employing a joint bilateral filtering method, the matching accuracy is further improved and noise interference is reduced.

[0043] 3) Parallax calculation

[0044] Based on the cost aggregation results, a dynamic programming algorithm is used to calculate the disparity of the same object in the left and right images, generating a disparity map.

[0045] 4) Parallax purification and smoothing

[0046] The disparity map is refined and noise smoothed using methods such as joint bilateral filtering to remove noise and repair void areas, generating a more accurate depth map. Oscillation signals are used to supplement and verify terrain information in the depth map, enhancing the robot's understanding and analysis of complex terrain.

[0047] (4) Weed area extraction and multidimensional feature analysis

[0048] Based on the generated depth map, the system identifies and extracts weed areas on the ground through threshold segmentation and the YOLOv9 algorithm based on deep learning. By analyzing multi-dimensional features such as shape, color, and texture, the system provides a basis for the controller to adjust the lifting components. By fusing image data and oscillation signals, the robot adjusts its navigation strategy and control parameters. The system analyzes terrain features in real time and optimizes the navigation path to achieve accurate classification and stable navigation in complex terrains.

[0049] A controller is mounted on the robot's outer shell 1. The controller is connected to the infrared sensor 4, the binocular camera 5, the walking motor, and the steering mechanism. When the infrared sensor 4 detects an obstacle, it sends a signal to the controller. The controller then controls the steering mechanism to deflect the steering wheels 2, enabling the robot to turn and avoid obstacles. The ground images captured by the binocular camera 5 are uploaded to the controller. The controller has a built-in learning algorithm that determines whether there are weeds on the ground, as well as the height and size of the weeds.

[0050] like Figure 1 , Figure 2 As shown, a lifting plate 6 is located at the lower front end of the robot shell 1. A lifting assembly for adjusting the height of the lifting plate 6 is located between the lifting plate 6 and the robot shell 1. The lifting assembly adjusts the height of the lifting plate 6 according to the height of the weeds. Figure 3 As shown, a lawnmower 7, which performs weeding by rotation, is rotatably mounted on the lifting plate 6 via the lawnmower shaft 21. The controller is also connected to the lifting assembly via a signal. When the area of ​​weeds on the ground exceeds a specified threshold, the controller sends a signal to the lifting assembly based on the depth map information provided by the binocular camera 5, causing the lifting assembly to adjust the height of the lifting plate 6 so that the height of the lawnmower 7 adapts to the height of the weeds. At the same time, the controller controls the lawnmower 7 to rotate to achieve weeding.

[0051] Specifically, such as Figure 3 As shown, the lifting assembly includes two symmetrically arranged telescopic cylinders 8. The cylinder barrels of the telescopic cylinders 8 are fixedly connected to the bottom of the robot shell 1, and the piston rods of the telescopic cylinders 8 are fixedly connected to the lifting plate 6. When the piston rods of the telescopic cylinders 8 extend or retract, the height of the lifting plate 6 is adjusted. When the telescopic cylinders 8 receive a signal from the controller, the piston rods of the telescopic cylinders 8 extend or retract. The controller sends a signal based on the depth map information from the binocular camera 5, causing the piston rods of the telescopic cylinders 8 to extend or retract, ensuring that the lawnmower 7 can adapt to the height of the weeds on the ground.

[0052] like Figure 2 , Figure 3As shown, there are two lawnmowers 7 symmetrically arranged on the lifting plate 6. The arrangement of two lawnmowers 7 can increase the weeding range, and the two lawnmowers 7 work together so that the weeds thrown out by one lawnmower 7 can be cut a second time by the other lawnmower 7, improving the weeding effect. The lawnmower 7 includes a blade disc 71 and blades 72 fixed to the outer wall of the blade disc 71. The blade disc 71 is fixed to the lower end of the lawnmower shaft 21, and the blades 72 are evenly distributed on the outer wall of the blade disc 71.

[0053] To drive the rotation of the lawnmower 7, such as Figure 3 As shown, a weeding motor 19 is fixed on the lifting plate 6, as... Figure 4 As shown, a drive bevel gear 11 is fixed on the output shaft 20 of the weeding motor 19. The blade 71 of the mower 7 is coaxially arranged with the driven bevel gear 10, that is, the driven bevel gear 10 is fixed to the lower end of the mower shaft 21, and the driven bevel gear 10 meshes with the drive bevel gear 11. A protective cover 9 is fixed on the lifting plate 6, covering the weeding motor 19, the drive bevel gear 11, and the driven bevel gear 10. The protective cover 9 protects the weeding motor 19, the drive bevel gear 11, and the driven bevel gear 10. The weeding motor 19, the drive bevel gear 11, and the driven bevel gear 10 constitute the drive mechanism, which is used to drive the rotation of the mower 7.

[0054] To achieve the shredding of cut weeds, such as Figure 5 As shown, the robot's outer shell 1 has a negative pressure pump 12 in the middle and a storage chamber 13 at the rear. The negative pressure pump 12 is connected to the storage chamber 13 via a pipe 22. Under the action of the negative pressure pump 12, the weeds cut by the lawnmower 7 are adsorbed into the storage chamber 13. The negative pressure pump 12, the pipe 22, and the storage chamber 13 constitute an adsorption assembly for adsorbing weeds.

[0055] like Figure 5 As shown, storage compartment 13 contains a crushing component, such as... Figure 6 As shown, the crushing assembly includes two crushing rollers 15 rotating in opposite directions. Each crushing roller 15 has crushing teeth 23 on its outer wall and is rotatably mounted inside the storage chamber 13. The bottom of the robot housing 1 has a discharge hole 14 communicating with the storage chamber 13 for discharging the crushed weeds. The two rotating crushing rollers 15 work together to squeeze and collide with the weeds, thereby crushing them. The crushed weeds are then discharged from the storage chamber through the discharge hole 14 under negative pressure.

[0056] like Figure 6As shown, a drive gear 16 is fixed on the shaft of one crushing roller 15, and a driven gear 17 is fixed on the shaft of the other crushing roller 15. The drive gear 16 and the driven gear 17 mesh and are located outside the robot housing 1. The driven gear 16 rotates synchronously with the drive wheel 18. The arrangement of the drive gear 16 and the driven gear 17, and their assembly relationship, achieve the purpose of the two crushing rollers 15 rotating in opposite directions. To drive the rotation of the crushing rollers 15, the drive gear 16 is coaxially arranged with the friction roller 18. The friction roller 18 contacts and presses against the drive wheel 3. When the drive wheel 3 rotates, it drives the friction roller 18 to rotate, which in turn drives the drive gear 16 to rotate, which in turn drives the driven gear 17 to rotate in the opposite direction, thus making the two crushing rollers 15 rotate synchronously but in opposite directions.

[0057] To prevent the lawnmower 7 from colliding hard with the ground and to protect the blades, such as... Figure 7 As shown, the lawnmower 7 includes a blade disc 71 and blades 72 fixed to the outer wall of the blade disc 71. The blades 72 have an arc-shaped structure with their protrusions facing downwards. The lower end of the lawnmower shaft 21 extends into a groove 73 at the lower end of the blade disc 71. The lower end of the lawnmower shaft 21 has a sliding plate 24, which is slidably connected to the groove 73 in the blade disc 71 in a vertical direction. A spring 25 is located between the bottom of the sliding plate 24 and the groove 73. When the spring 25 is at its natural length, the height of the blades 72 is at its lowest. When the blade disc 71 contacts the ground and the ground exerts an upward squeezing force on the blade disc 71, the spring 25 is compressed, causing the blade disc 71 to move upwards a certain distance. This prevents the blade disc 71 from making hard contact with the ground, thus preventing the blades 72 from being damaged by hard contact with the ground and improving the service life of the lawnmower 7. The blade 72 has an arc-shaped structure with its protruding part facing downwards. This protruding part contacts the weeds for weed removal. The tip of the blade 72 is raised to prevent it from digging into the soil and causing damage, and it adapts to uneven ground, allowing it to rise and fall with the terrain to meet the needs of weed removal at different heights. The lifting assembly provides coarse adjustment of the mower 7's height, while the spring 25 provides fine adjustment. This combination of coarse and fine adjustment allows for precise and gentle adjustment of the mower 7's height.

[0058] The present invention discloses a method for operating a weeding robot in hilly and mountainous areas, comprising the following steps:

[0059] S1. The robot autonomously navigates along a set route in the orchard and uses infrared sensor 4 to scan and measure the time difference and phase difference of transmitted and received signals to monitor the specific location and distance of obstacles in real time.

[0060] S2. High-precision acquisition of orchard images is achieved through binocular camera 5. A stereo matching algorithm is used to register and process the left and right images, generating a precise depth map of the orchard scene. Based on image preprocessing and target detection technology, the extracted weed area images are analyzed in depth, multi-dimensional weed feature extraction is performed, and a complete weed feature set is constructed. When the detected ground weed area exceeds a specified threshold, the controller intelligently adjusts the height of the lifting plate 6 based on the depth map information provided by binocular camera 5, and the lawnmower 7 rotates to clear the ground weeds.

[0061] S3. The robot uses a multi-input, multi-output vibration model for terrain classification. Using an oscillation sensor and a binocular camera 5 mounted on the robot, oscillation signals and image data are acquired in real time. The oscillation signals are used to supplement and verify terrain information in the image data. When the robot travels through complex terrain, it adjusts its navigation strategy and control parameters based on the fusion results of the image data and oscillation signals to achieve accurate terrain classification and stable navigation.

[0062] S4. Start the negative pressure pump 12 to suck the cut weeds into the storage bin 13, and crush the weeds through the crushing roller 15. The crushed weeds are discharged from the storage bin 13 through the discharge hole 14.

[0063] This invention uses an infrared sensor 4 to detect obstacles in front of the robot, enabling timely obstacle avoidance. A binocular camera 5 acquires images of the robot's front and the ground, and uses a stereo matching algorithm to accurately determine whether weeds exceed a specified threshold. The oscillation signal and image data are fused to adjust the navigation strategy in real time. The lifting assembly intelligently adjusts the height of the lifting plate 6 to adapt to the weed removal needs of different heights. The lawnmower 7 quickly cuts ground weeds through rotation, achieving efficient removal. The negative pressure pump 12 sucks the cut weeds into the storage chamber 13, where they are then crushed by a crushing assembly and discharged through the discharge hole 14. The crushing assembly is designed not only to handle cut weeds but also to crush incompletely cut roots, preventing weed regrowth. The elastic connection between the lawnmower 7 and the lawnmower shaft 21 allows for fine-tuning of the height, bringing the blade 72 closer to the weed roots, further improving the removal effect and reducing the possibility of weed regrowth.

[0064] It should be noted that the structures, proportions, sizes, etc., illustrated in this invention are only for illustrating the technical content disclosed in the specification and are not intended to limit the implementation conditions of this invention. Any modifications to the structure, changes in the proportional relationships, or adjustments to the size, as long as they do not affect the effects and technical objectives achieved by this invention, should still be within the technical scope disclosed in this invention. Reasonable adjustments to the implementation details of this invention can be made to adapt to different application needs and technical environments without affecting the core innovation and technical effects of this invention.

[0065] In the description of this invention, terms such as "upper," "lower," "front," "rear," "left," and "right" are primarily used for clarity of description and not for limiting the scope of the invention. The above directional descriptions are relative and, depending on the specific installation and usage environment during actual implementation, are still considered to fall within the scope of the invention without substantially altering the technical content.

[0066] In this invention, unless otherwise explicitly specified and limited, terms such as "connection" and "fixation" should be interpreted broadly. For example, a connection can be a fixed connection (such as welding, bonding, etc.), a detachable connection (such as threaded connection, snap-fit ​​connection, etc.), or an integral connection (such as a molded integral component). Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

Claims

1. A method for operating a hilly and mountainous weeding robot, the hilly and mountainous weeding robot comprising a robot shell, a steering wheel at the front end of the robot shell, and a drive wheel at the rear end of the robot shell, characterized in that, The steering wheel and drive wheel are equipped with oscillation sensors for detecting oscillation signals. The front sidewall of the robot shell is equipped with an infrared sensor for detecting obstacles, and a binocular camera for capturing ground images to determine the presence and height of weeds. A lifting plate is located below the front end of the robot shell, and a lifting assembly for adjusting the height of the lifting plate is located between the lifting plate and the robot shell. The lifting assembly adjusts the height of the lifting plate according to the height of the weeds. A weed cutter is rotatably mounted on the lifting plate via a weed cutter shaft. A negative pressure pump is located in the middle of the robot shell, and a storage chamber is located at the rear end of the robot shell. The negative pressure pump is connected to the storage chamber via a pipe. A crushing assembly is located inside the storage chamber. The crushing assembly includes two crushing rollers rotating in opposite directions. The outer wall of the crushing rollers has crushing teeth, and the crushing rollers are rotatably mounted inside the storage chamber. The bottom of the robot shell has a discharge hole connected to the storage chamber for discharging the crushed weeds. The robot shell is characterized by the following steps: S1. The robot autonomously navigates along a set route in the orchard and uses the infrared sensor to dynamically scan and measure the time difference and phase difference of transmitted and received signals to monitor the specific location and distance of obstacles in real time. S2. High-precision acquisition of orchard images is achieved through the binocular camera. The left and right images are registered and processed using a stereo matching algorithm to generate a precise depth map of the orchard scene. Based on image preprocessing and target detection technology, the extracted weed area images are analyzed in depth to extract multi-dimensional weed features and construct a complete weed feature set. When the detected ground weed area exceeds a specified threshold, the height of the lifting plate is intelligently adjusted according to the depth map information provided by the binocular camera, and the lawnmower is started to rotate to clear the ground weeds. S3. The robot includes a multi-input, multi-output three-degree-of-freedom vibration model, which contains four input signals and three output signals. During the robot's movement, the controller extracts the four input signals from the oscillation sensors connected to the robot's steering wheels and drive wheels. The output signals include the pitch rate, roll rate, and vertical rate captured by the robot's speed sensor, used to quantify the robot's dynamic behavior and response during movement, describing the vehicle's rotational speed around the horizontal and vertical axes and the acceleration change in the vertical direction. The robot uses the multi-input, multi-output vibration model for terrain classification. Through the oscillation sensors and binocular cameras mounted on the robot, oscillation signals and image data are collected in real time. The oscillation signals are used to supplement and verify the terrain information in the image data. When the robot is driving in complex terrain, the navigation strategy and control parameters are adjusted according to the fusion result of the image data and oscillation signals to achieve accurate terrain classification and stable navigation. Based on the generated depth map, the ground weed area is identified and extracted through threshold segmentation and the YOLOV algorithm based on deep learning. By analyzing the multi-dimensional features of shape, color, and texture, the controller provides a basis for adjusting the lifting components. S4. Start the negative pressure pump to suck the cut weeds into the storage bin, crush the weeds through the crushing component, and discharge the crushed weeds through the discharge hole into the storage bin.

2. The working method of a hilly and mountainous weeding robot according to claim 1, characterized in that, The lifting assembly includes two symmetrically arranged telescopic cylinders. The cylinder barrels of the telescopic cylinders are fixedly connected to the robot's outer shell, and the piston rods of the telescopic cylinders are fixedly connected to the lifting plate.

3. The working method of a hilly and mountainous weeding robot according to claim 1, characterized in that, The lawnmower consists of two symmetrically arranged devices on the lifting plate.

4. The working method of a hilly and mountainous weeding robot according to claim 3, characterized in that, A weeding motor is fixed on the lifting plate, and a driving bevel gear is fixed on the output shaft of the weeding motor. The weed cutter is coaxially arranged with the driven bevel gear, and the driven bevel gear meshes with the driving bevel gear. A protective cover is fixed on the lifting plate to cover the weeding motor, the driving bevel gear and the driven bevel gear.

5. The working method of a hilly and mountainous weeding robot according to claim 1, characterized in that, One of the crushing rollers has a drive gear fixed on its roller shaft, and the other crushing roller has a driven gear fixed on its roller shaft. The drive gear and the driven gear mesh and are located outside the robot's outer shell. The drive gear rotates synchronously with the drive wheel.

6. The working method of a hilly and mountainous weeding robot according to claim 5, characterized in that, The drive gear is coaxially arranged with the friction roller, the friction roller is in contact and pressed with the drive wheel, and the drive wheel rotates to drive the friction roller to rotate.

7. The working method of a hilly and mountainous weeding robot according to claim 1, characterized in that, The lawnmower includes a blade disc and blades fixed to the outer wall of the blade disc. The blades are arc-shaped blades with their protruding parts facing downwards. The lower end of the lawnmower shaft extends into the blade disc. The lower end of the lawnmower shaft has a sliding plate. The sliding plate is slidably connected to a groove in the blade disc and slides in the up-down direction. A spring is located between the bottom of the sliding plate and the groove. The blade height is lowest when the spring is at its natural length.

Citation Information

Patent Citations

  • Intelligent weeding robot

    CN112567960A

  • Intelligent mower based on laser scanning radar sensor and control method of intelligent mower

    CN108781705A

  • Vibration information terrain classification and identification method based on CNN-LSTM

    CN110956154A

  • Agricultural weeding machine based on new energy

    CN111837594A

  • Weeding device for landscaping

    CN219228422U