Solar laser weeding machine and weeding method

Abstract

The invention provides a solar laser weeding machine and a weeding method.The weeding machine comprises a guiding device installed in the center of the front of a chassis, the guiding device senses rice seedlings through pressure sensors on the two sides and assists in judging whether the weeding machine walks in the middle of a rice field or not, and then guiding control is conducted; wheels are installed on the two sides of the chassis, an integrated computer, a storage battery and a platform are installed on the chassis, a rotating mechanism is installed on the platform, and a laser sensor is installed on the rotating mechanism. The weeding machine is powered by solar energy, an inter-plant weeding mechanism is installed below the chassis and close to the tail, and an inter-row weeding mechanism is installed in a protective cover at the tail of the chassis. The integrated computer obtains three-dimensional point cloud data of plants based on rotary scanning of the laser sensor, rice growing in rows and randomly distributed weeds can be efficiently distinguished in combination with a DBSCAN density clustering algorithm, and meanwhile stable operation of equipment under various topographic conditions is ensured through the innovative suspension system and pressure sensor design.

Description

Technical Field

[0001] This invention belongs to the field of weeding machine design technology, and particularly relates to a solar-powered laser weeding machine and weeding method. Background Technology

[0002] In recent years, the rice planting area in my country has been increasing. Among them, weeding is related to the rice yield. The main methods of weeding in rice fields include manual weeding, chemical weeding, and mechanical weeding.

[0003] Manual weeding is economical, environmentally friendly, and safe, but it is labor-intensive and inefficient. With urbanization and an aging population, manual weeding faces a labor shortage. Chemical weeding is economical and labor-saving, but long-term, uncontrolled use can lead to weed resistance, crop damage, and environmental and water pollution. Ordinary mechanical weeding easily damages seedlings; small spacing between plants and between nearby plants, coupled with intertwined root systems, makes operation difficult and damages root systems, resulting in a high rate of seedling damage.

[0004] There is currently some research on laser weeding machines. Existing technology (CN202411039725.1) proposes a laser weeding device based on visual recognition, which uses a camera to acquire images and then uses lasers for weeding. While this significantly reduces the probability of damaging seedlings, the reliance on visual weed recognition leads to a long training cycle, high equipment costs, complex image processing algorithms, and poor real-time performance, making it difficult to meet the needs of large-scale field operations. Furthermore, existing laser weeding machines generally lack effective terrain adaptation mechanisms, easily resulting in poor weeding effects in hilly, mountainous, and winding terrain. Additionally, traditional laser weeding machines heavily rely on structured agricultural scenarios when processing point cloud data; they cannot adapt to irregular farmland environments. They also heavily depend on external auxiliary information, pre-trained data, and preset scanning height differences, making it impossible to independently distinguish between rice and weeds. They are extremely sensitive to dynamic environments; wind rustling and obstacles can severely affect recognition results, demonstrating poor robustness to dynamic conditions.

[0005] To solve the above-mentioned technical problems, this invention designs a novel solar-powered laser weeding machine and weeding method. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a solar-powered laser weeding machine and weeding method. Through an innovative laser scanning mechanism and intelligent recognition algorithm, it solves the problems of scanning blind spots and poor environmental adaptability of existing equipment. The two-dimensional laser sensor of this invention, combined with a rotating mechanism, realizes low-cost three-dimensional scanning function. Combined with the DBSCAN density clustering algorithm, it significantly improves the accuracy of weed identification. At the same time, the innovative suspension system and pressure sensor design ensure stable operation of the equipment under various terrain conditions.

[0007] The present invention achieves the above-mentioned technical objectives through the following technical means.

[0008] A solar-powered laser weeding machine includes a chassis, a guide device installed at the center of the front of the chassis, wheels installed on both sides of the chassis, an integrated computer, a battery, and a platform installed on the chassis, a rotating mechanism installed on the platform, and a laser sensor installed on the rotating mechanism; a solar panel is installed above the battery; an inter-row weeding mechanism is installed below the chassis near the rear; and a protective cover is also installed at the rear of the chassis, inside which the inter-row weeding mechanism is installed.

[0009] Furthermore, the guiding device includes a housing, a locking pin, and a flexible rod; the locking pin is installed at the left and right openings of the housing to fix the flexible rod, and multiple pressure sensors are distributed on the front side of the flexible rod; when the pressure sensors come into contact with the rice seedlings, they sense the pressure, and the microcontroller built into the housing calculates the difference between the distance from the pressure sensor closest to the housing that senses the pressure to the center line of the housing in the pressure sensor array on the left and right sides to determine whether the weeder is in the middle of the rice field. When the difference is within the range of [-1cm, 1cm], it is determined that it is in the middle of the rice field; otherwise, it indicates that the weeder has deviated, and the microcontroller then controls the steering of the weeder.

[0010] Furthermore, the inter-plant weeding mechanism includes multiple Y-shaped fixing frames installed near the rear of the chassis. Each fixing frame has an inter-plant weeding wheel fixed at both ends via a rotating shaft. Each inter-plant weeding wheel consists of three blades with adjacent angles of 120°, and all three blades are fixed on the same rotating shaft. The other ends of the two rotating shafts extend into the fixing frame, and universal joints are installed at the ends of the rotating shafts. Gears A driven by motor C are installed inside the fixing frame. Both gears A mesh with corresponding driven gears B. The driven gears B have elongated holes. Bearings are also installed on the rotating shaft inside the fixing frame. A connecting cylinder is fixed on the bearing. The lower end of the connecting cylinder passes through the driven gear B and is fixed therein.

[0011] Furthermore, the inter-row weeding mechanism includes a fixed...

[0012] Motor B, located at one end of the protective cover, is connected to the power output shaft via a coupling. The other end of the power output shaft is rotatably mounted on the other end of the protective cover. Multiple inter-row weeding wheels are fixed on the power output shaft. Each inter-row weeding wheel includes three evenly distributed blades, with the end of each blade bent outwards.

[0013] Furthermore, the rotating mechanism includes a rotating platform, a gearbox, a motor A, a driven gear A, and a gear B. Motor A and the gearbox are both fixed to the platform. Gear B is fixed to the power output shaft of motor A. Gear B meshes with the driven gear A inside the gearbox, enabling motor A to drive gear B to rotate. Gear B further meshes with the driven gear A to rotate. A first connecting shaft is fixed to the center of driven gear A. The portion of the first connecting shaft extending out of the gearbox is fixed to the rotating platform. A laser sensor is fixed to the rotating platform via a second connecting shaft. Driven gear A drives the rotating platform to rotate together, thereby causing the laser sensor on it to rotate and collect information.

[0014] A weeding method using the aforementioned solar-powered laser weeding machine includes the following steps:

[0015] Step 1: Start the weeding machine and initialize all equipment;

[0016] Step 2: As the weeder moves across the field, the laser sensor emits a laser beam and measures the time it takes for the laser to reflect back, acquiring distance information about the object's surface. Combined with a rotating platform, it performs a full-range scan, collecting a large amount of point data, which is then transmitted to an integrated computer for processing and analysis, generating a 3D point cloud dataset of the plants. This dataset contains the 3D coordinates of each point. Then, based on the laser sensor's measurement accuracy threshold, the point cloud data undergoes preliminary processing to obtain a cleaned point cloud dataset.

[0017] Step 3: The integrated computer uses the DBSCAN algorithm to further identify the point cloud dataset. Referring to the distribution characteristics of rice and weed point clouds, a unified clustering parameter is manually preset: neighborhood radius. =0.08m, minimum number of core points =4; then for any point in the point cloud dataset obtained in step 2... ,That Neighborhood is denoted as , ,in For point With point Euclidean distance, If point of The number of points contained in the neighborhood is not less than ,Right now Then point As the core point;

[0018] If there exists a sequence of points , , ..., For any , All are core points and ,but from Density can reach, denoted as ;

[0019] If a cluster Satisfies: For any two points within a cluster, there exists a common point within the cluster such that their mutual density is achievable (i.e., for any...). , ,exist ,make and Points within a cluster and points outside the cluster have no density reachability relationship (i.e., for points within the cluster, there is no density reachability relationship). and , and If an isolated point does not meet the density reachability requirement, it is considered a noise point, and this cluster is then classified as part of the same cluster.

[0020] During the movement of the weeding machine, each time new point cloud data is collected, the new data is merged with the historical point cloud data into a full dataset, and the DBSCAN clustering process described above is re-executed. The neighborhood, core points and cluster relationships are repeatedly calculated for the full dataset. Based on the clustering results and combined with the distribution pattern of rice, densely distributed clusters are identified as rice, and sparsely distributed or scattered small clusters are identified as weeds, thus obtaining the three-dimensional coordinate positions of rice and weeds.

[0021] Step 4: As the weeder moves forward, when the pressure sensor installed at the front end of the flexible rod comes into contact with the rice seedling and is deformed by the pressure of the rice plant, it sends a pressure signal back to the microcontroller inside the guide device. The microcontroller calculates the difference and determines whether the weeder has deviated from the center of the rice row, and then controls the wheels to correct the path.

[0022] Meanwhile, the integrated computer transmits the rice and weed location data obtained in step 3 to its internal path planning module and weeding control module; the path planning module, combined with the rice row centerline position sensed by the pressure sensor, plans the weeder's travel path to ensure that it avoids the rice cluster coverage area and covers the weed cluster distribution area; the weeding control module controls the weeding wheel between rows to rotate continuously as the weeder moves forward, and the curved design at the end of the blade helps to break up the soil and remove weeds between rows;

[0023] When the clustering results show that there are weed clusters between plants, the integrated computer sends a command to motor C to drive gear A to rotate. By controlling the direction and angle of rotation of gear A, the rotation of driven gear B is controlled. Driven gear B drives the shaft to swing, which in turn drives the weeding wheels between plants to rotate closer to each other and cut the weed clusters between plants. The rice passes through the gap between the two weeding wheels to avoid damaging the seedlings.

[0024] The present invention has the following beneficial effects:

[0025] The dual precision identification of laser and algorithm: The two-dimensional laser sensor obtains three-dimensional point cloud data of plants through rotational scanning. Combined with the DBSCAN density clustering algorithm, based on the regular distribution of rice with "row spacing of 30cm and plant spacing of 10-15cm", it can efficiently distinguish between rice growing in rows and randomly distributed weeds.

[0026] Single-line laser sensors can typically achieve high accuracy in measuring rice plant height, with errors generally controlled within a few millimeters, enabling them to accurately acquire rice height information. By utilizing the differences in height and growth patterns between rice and weeds to set thresholds, false positives can be further filtered out, improving accuracy.

[0027] By combining a two-dimensional laser sensor with a rotating mechanism, the function of a three-dimensional laser sensor was realized, which greatly reduced the cost.

[0028] Two rows of pressure sensors detect voltage changes when the rice plants come into contact in real time, calculate the distance the weeder deviates from the center line of the rice row, and dynamically adjust the direction of travel to avoid misjudgment caused by mechanical transplanting deviation or terrain undulation, thus ensuring the accuracy of the weeding path.

[0029] The inter-row weeding blade has a linkage structure that can automatically adapt to changes in terrain to carry out inter-row weeding operations. It combines environmental friendliness, precision, adaptability, high efficiency and multi-functional integration, realizing automatic weeding in paddy fields. Attached Figure Description

[0030] Figure 1 This is a schematic diagram of the overall structure of the solar-powered laser weeding machine described in this invention;

[0031] Figure 2 This is a schematic diagram showing the disassembled structure of the guiding device;

[0032] Figure 3 This is a schematic diagram of the internal structure of the rotating platform and gearbox.

[0033] Figure 4 This is a schematic diagram of the tail-end weeding blade;

[0034] Figure 5 A schematic diagram of the installation of weeding wheels between plants;

[0035] Figure 6 This is a schematic diagram showing the connection between gear A, driven gear B, and rotating shaft inside the fixed frame.

[0036] In the diagram: 1. Housing; 2. Locking pin; 3. Flexible rod; 4. Pressure sensor; 5. Integrated computer; 6. Rotating platform; 7. Laser sensor; 8. Solar panel; 9. Platform; 10. Chassis; 11. Gearbox; 12. Battery; 13. Gear A; 14. Wheel; 15. Motor A; 16. Inter-row weeding wheel; 17. Inter-plant weeding wheel; 18. Shaft; 19. Driven gear A; 20. Gear B; 21. Power take-off shaft; 22. Motor B; 23. Fixture; 24. Driven gear B. Detailed Implementation

[0037] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but the scope of protection of the present invention is not limited thereto.

[0038] To facilitate understanding of the present invention, it should be understood in the following description that the terms "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. In the description of the present invention, it should be noted that, unless otherwise expressly specified and limited, the terms "installed," "connected," and "linked" should be interpreted broadly. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, an indirect connection through an intermediate medium, or a connection within two components.

[0039] like Figure 1 As shown, the solar-powered laser weed cutter of the present invention includes a guiding device, a chassis 10, a laser sensor 7, a rotating mechanism, a battery 12, and a solar panel 8.

[0040] like Figure 1 , 2As shown, the guide device is installed at the front center of the chassis 10, and includes a housing 1, a locking pin 2, and a flexible rod 3. The locking pin 2 is installed at the left and right openings of the housing 1 to fix the flexible rod 3 and prevent water from entering. Small pressure sensors 4 are distributed on the front side of the flexible rod 3. During the movement of the weeder, when the pressure sensors 4 come into contact with the rice seedlings, some of the pressure sensors 4 sense the pressure. The microcontroller built into the housing 1 calculates the difference between the distances of the pressure sensors 4 that are closest to the housing 1 and sense the pressure to the center line of the housing 1 (in this embodiment, the preferred difference is the distance from the right pressure sensor 4 to the center line minus the distance from the left pressure sensor 4 to the center line) to determine whether the weeder is moving in the middle of the rice field. When the difference is within the range of [-1cm, 1cm], it is determined that the weeder is moving in the middle of the rice field. When the difference is greater than 1cm, the microcontroller issues a command to control the weeder to turn right. When the difference is less than -1cm, the microcontroller controls the weeder to turn left. As the machine moves forward, the flexible rod 3 passes over the rice seedlings and automatically returns to its original shape using its own elasticity, before proceeding to the next judgment.

[0041] like Figure 1 As shown, wheels 14 are mounted on both sides of the chassis 10. An integrated computer 5, a battery 12, and a platform 9 are mounted on the chassis 10. A rotating mechanism is mounted on the platform 9. The rotating mechanism is controlled by the integrated computer 5 to rotate back and forth. A laser sensor 7 is mounted on the rotating mechanism. The single-line laser sensor 7 can only acquire distance data on a two-dimensional plane. This invention adds a rotating mechanism, which enables it to drive the laser sensor 7 to rotate and collect information, ensuring high-speed and stable rotational scanning.

[0042] like Figure 1 , 3 As shown, the rotating mechanism is essentially a two-axis turntable structure, including a rotating platform 6, a gearbox 11, a motor A15, a driven gear A19, and a gear B20. The motor A15 and gearbox 11 are both fixed on the platform 9. The gear B20 is fixed on the power output shaft of the motor A15. The gear B20 meshes with the driven gear A19 inside the gearbox 11, and the motor A15 can drive the gear B20 to rotate. The gear B20 further meshes with the driven gear A19 to rotate. A first connecting shaft is fixed at the center of the driven gear A19. The part of the first connecting shaft that extends out of the gearbox 11 is fixed to the rotating platform 6. A laser sensor 7 is fixed on the rotating platform 6 through a second connecting shaft. The driven gear A19 can drive the rotating platform 6 to rotate together, thereby driving the laser sensor 7 on it to rotate and collect information.

[0043] like Figure 1 As shown, the lawnmower is powered by solar energy. A solar panel 8 is installed above the battery 12 and is connected to the battery 12 to store electrical energy.

[0044] likeFigure 1 , 5 As shown in Figure 6, multiple Y-shaped mounting brackets 23 are installed near the rear of the weeder chassis 10. Each mounting bracket 23 has a weeding wheel 17 fixed to both ends via a rotating shaft 18. Specifically, each weeding wheel 17 consists of three blades with adjacent angles of 120°, and all three blades are fixed to the same rotating shaft 18. The other ends of both rotating shafts 18 extend into the mounting bracket 23, and universal joints are installed at the ends of the rotating shafts 18. Gears A13 driven by a motor C are installed inside the mounting bracket 23. Both gears A13 mesh with corresponding driven gears B24 for transmission. The driven gear B24 has an elongated hole. A bearing is installed on the rotating shaft 18 inside the fixed frame 23. A connecting cylinder is fixed to the bearing, and its lower end passes through and is fixed to the driven gear B24. Motor C drives gear A13 to rotate. By controlling the direction and angle of rotation of gear A13, the rotation of driven gear B24 is controlled. Driven gear B24 drives the oscillation of rotating shaft 18, thus achieving three degrees of freedom for the rotation and oscillation of shaft 18. Finally, rotating shaft 18 drives the two weeding wheels 17 to rotate, bringing them closer together for weeding. In practical applications, when the integrated computer 5 detects a clump of weeds between plants, it sends a command to motor C. Motor C controls the rotation direction and angle of gear A13 to drive rotating shaft 18 towards each other, bringing the two weeding wheels 17 closer together for precise cutting of the weed clumps. Rice plants pass through the gap between the two weeding wheels 17 without damage.

[0045] like Figure 1 , 4 As shown, a protective cover is also installed at the rear of the weeder chassis 10. Inside the protective cover is an inter-row weeding mechanism. The inter-row weeding mechanism includes a motor B22 fixed at one end of the protective cover. The inter-row weeding wheel drive motor B22 is connected to a power output shaft 21 via a coupling. The other end of the power output shaft 21 is rotatably mounted on the other end of the protective cover. Multiple inter-row weeding wheels 16 are fixed on the power output shaft 21. Each inter-row weeding wheel 16 includes three evenly distributed blades, each blade tip of which is bent outward to facilitate breaking up soil and weeds. During operation, the inter-row weeding wheels 16 move with the machine and rely on continuous rotation to break up the soil and remove the inter-row weeds.

[0046] The weeding method using the aforementioned solar-powered laser weed cutter includes the following steps:

[0047] Step 1: Start the weed cutter, initialize the integrated computer 5, and check whether the laser sensor 7 and pressure sensor 4 are working properly. Confirm that the laser sensor 7 is not obstructed and the rotating mechanism is not stuck.

[0048] Step 2: The weeder moves across the field. The laser sensor 7 emits a laser beam and measures the time it takes for the laser to reflect back, acquiring distance information about the object's surface. This information is then used in conjunction with the rotating platform 6 for omnidirectional scanning. The rotating mechanism is driven by motor A15 and achieves stable rotation through gearbox 11, thereby collecting a large amount of point data and transmitting it to the integrated computer 5. The integrated computer 5 processes and analyzes this point data to generate a three-dimensional point cloud dataset of the plants. The dataset contains the three-dimensional coordinate data of each point. Then, based on the measurement accuracy threshold of the laser sensor 7, the point cloud data is preliminarily processed to remove points whose distance values ​​exceed the normal operating range of 0.1 to 5 meters, and to filter out isolated noise points caused by equipment errors and environmental interference, resulting in a purified point cloud dataset.

[0049] Step 3: Rice plants generally have a relatively uniform height, while weeds vary in height. Furthermore, rice plants are typically planted in rows and columns, exhibiting a certain regularity. Therefore, the integrated computer 5 utilizes a density-based spatial clustering algorithm (DBSCAN) for further identification. Referring to the point cloud distribution characteristics of rice and weeds, a unified clustering parameter is manually preset: neighborhood radius. =0.08m, minimum number of core points =4; then for any point in the point cloud dataset obtained in step 2... ,That Neighborhood is denoted as , ,in For point With point Euclidean distance, If point of The number of points contained in the neighborhood is not less than ,Right now Then point As the core point;

[0050] If there exists a sequence of points , , ..., For any , All are core points and ,but from Density can reach, denoted as ;

[0051] If a cluster Satisfies: For any two points within a cluster, there exists a common point within the cluster such that their mutual density is achievable (i.e., for any...). , ,exist ,make and Points within a cluster and points outside the cluster have no density reachability relationship (i.e., for points within the cluster, there is no density reachability relationship). and , and Isolated points that do not meet the density reachability requirement are considered noise points. Therefore, this cluster is classified as part of the same cluster.

[0052] During the movement of the weeding machine, each time new point cloud data is collected, the new data is merged with the historical point cloud data into a full dataset, and the DBSCAN clustering process described above is re-executed. There is no incremental processing logic; the neighborhood, core points, and cluster relationships are repeatedly calculated for the full dataset. Based on the clustering results and the distribution pattern of rice with "row spacing of 30cm and plant spacing of 10-15cm", densely distributed clusters (with many points within each cluster) are identified as rice, while sparsely distributed clusters (with few points within each cluster) or scattered small clusters are identified as weeds, thus obtaining the three-dimensional coordinate positions of rice and weeds.

[0053] Step 4: As the weeder moves forward, when the pressure sensor 4 installed at the front end of the flexible rod 3 comes into contact with the rice seedling and is deformed by the pressure of the rice plant, it sends a pressure signal back to the microcontroller inside the guide device. The microcontroller calculates the difference between the distance between the pressure sensors 4 on the left and right sides and the center line to determine whether the weeder has deviated from the center of the rice row. If the difference exceeds ±1 cm, the wheel is controlled to correct the path to ensure that the machine travels stably between the rice rows.

[0054] Meanwhile, the integrated computer 5 transmits the rice and weed location data obtained in step 3 to its internal path planning module and weeding control module; the path planning module, combined with the rice row centerline position sensed by the pressure sensor 4, plans the weeder's travel path to ensure that it avoids the rice cluster coverage area and covers the weed cluster distribution area; the weeding control module controls the inter-row weeding wheel 16 to rotate continuously as the weeder moves forward, and the curved design at the end of the blade helps to break up the soil and remove the inter-row weeds.

[0055] When the clustering results show that there are weed clusters between plants, the integrated computer 5 calculates the moment when the weeds between plants come into contact with the weeding wheel 17 based on the machine's running speed, sends a command to the motor C to drive the gear A13 to rotate, and controls the rotation of the driven gear B24 by controlling the direction and angle of the gear A13. The driven gear B24 drives the rotating shaft 18 to swing, which in turn drives the weeding wheels 17 between plants to rotate closer to each other, accurately cutting the location of the weed clusters between plants. The rice passes through the gap between the two weeding wheels, avoiding damage to the seedlings.

[0056] The embodiments described above are preferred embodiments of the present invention, but the present invention is not limited to the above embodiments. Any obvious improvements, substitutions or modifications that can be made by those skilled in the art without departing from the essence of the present invention shall fall within the protection scope of the present invention.

Claims

1. A solar-powered laser weeding machine, characterized in that, Includes a chassis (10), a guide device is installed at the front center of the chassis (10), wheels (14) are installed on both sides of the chassis (10), an integrated computer (5), a storage battery (12), and a platform (9) are installed on the chassis (10), a rotating mechanism is installed on the platform (9), and a laser sensor (7) is installed on the rotating mechanism; a solar panel (8) is installed above the storage battery (12); an inter-row weeding mechanism is installed below the chassis (10) near the rear; a protective cover is also installed at the rear of the chassis (10), and an inter-row weeding mechanism is installed inside the protective cover.

2. The solar-powered laser weed cutter according to claim 1, characterized in that, The guiding device includes a housing (1), a locking pin (2), and a flexible rod (3). The locking pin (2) is installed at the left and right openings of the housing (1) to fix the flexible rod (3). Multiple pressure sensors (4) are distributed on the front side of the flexible rod (3). When the pressure sensor (4) comes into contact with the rice seedling, it senses the pressure. The microcontroller built into the housing (1) calculates the difference between the distance from the pressure sensor (4) closest to the housing (1) in the pressure sensor array on the left and right sides that senses the pressure and the center line of the housing (1). This determines whether the weeder is walking in the middle of the rice field. When the difference is within the range of [-1cm, 1cm], it is determined that it is walking in the middle of the rice field. Otherwise, it indicates that the weeder has deviated. The microcontroller then controls the steering of the weeder.

3. The solar-powered laser weed cutter according to claim 2, characterized in that, The inter-plant weeding mechanism includes multiple Y-shaped fixing frames (23) installed near the tail of the chassis (10). Each fixing frame (23) has an inter-plant weeding wheel (17) fixed at both ends via a rotating shaft (18). Each inter-plant weeding wheel (17) consists of three blades with an adjacent angle of 120°, and all three blades are fixed on the same rotating shaft (18). The other ends of the two rotating shafts (18) extend into the fixing frame (23), and a universal joint is installed at the end of the rotating shaft (18). A gear A (13) driven by a motor C is installed inside the fixing frame (23). Both gears A (13) mesh with the corresponding driven gears B (24). The driven gears B (24) have elongated holes. A bearing is also installed on the rotating shaft (18) inside the fixing frame (23). A connecting cylinder is fixed on the bearing. The lower end of the connecting cylinder passes through the driven gear B (24) and is fixed therein.

4. The solar-powered laser weed cutter according to claim 3, characterized in that, The inter-row weeding mechanism includes a motor B (22) fixed at one end of the protective cover. The inter-row weeding wheel drive motor B (22) is connected to the power output shaft (21) via a coupling. The other end of the power output shaft (21) is rotatably mounted on the other end of the protective cover. Multiple inter-row weeding wheels (16) are fixed on the power output shaft (21). Each inter-row weeding wheel (16) includes three evenly distributed blades, each blade having an outwardly curved end.

5. The solar-powered laser weed cutter according to claim 1, characterized in that, The rotating mechanism includes a rotating platform (6), a gearbox (11), a motor A (15), a driven gear A (19), and a gear B (20). The motor A (15) and the gearbox (11) are both fixed on the platform (9). The gear B (20) is fixed on the power output shaft of the motor A (15). The gear B (20) meshes with the driven gear A (19) inside the gearbox (11). The motor A (15) can drive the gear B (20) to rotate. The gear B (20) further meshes with the driven gear A (19) to rotate. The center of the driven gear A (19) is fixed with a first connecting shaft. The part of the first connecting shaft that extends out of the gearbox (11) is fixed with the rotating platform (6). The rotating platform (6) is fixed with a laser sensor (7) through a second connecting shaft. The driven gear A (19) drives the rotating platform (6) to rotate together, thereby driving the laser sensor (7) on it to rotate and collect information.

6. A weeding method using the solar-powered laser weeding machine of claim 4, characterized in that, The process includes the following: Step 1: Start the weeding machine and initialize all equipment; Step 2: The weeder moves in the field. The laser sensor (7) emits a laser beam and measures the time it takes for the laser to reflect back to obtain distance information on the surface of the object. It is then used in conjunction with the rotating platform (6) to perform an all-round scan, thereby collecting a large amount of point data and transmitting it to the integrated computer (5) for processing and analysis to generate a three-dimensional point cloud dataset of the plants. Then, the point cloud data is preliminarily processed based on the measurement accuracy threshold of the laser sensor (7). Step 3: The integrated computer (5) uses the DBSCAN algorithm to further identify the point cloud dataset. Referring to the distribution characteristics of rice and weed point clouds, the unified clustering parameters are preset manually: neighborhood radius ε = 0.08m, minimum number of core points MinPts = 4. Then, for any point p in the point cloud dataset obtained in step 2, its ε-neighborhood is denoted as N. ε (p), N ε (p)={q∈D|dist(p,q)≤ε}, where dist(p,q) is the point p(x) p y p , z p ) and point q(x) q y q , z q The Euclidean distance of |N>; if |N> ε If (p)|≥MinPts, then point p is the core point; If there exists a point sequence p1, p2, ..., p... k ∈D, for any 1≤i≤k, p i Both are core points and p i+1 ∈N ε (p i If q can be reached from density p, then q is denoted as p→q; If a cluster C satisfies the following: any two points within the cluster have a common point within the cluster such that they are density-reachable from each other, i.e., for any p, q∈C, there exists r∈C such that r→p and r→q; and points within the cluster are not density-reachable from points outside the cluster, i.e., for any p ∈C and q ∈ C, there exists r ∈C such that r→p and r→q. If p and q do not satisfy density reachability, isolated points that do not satisfy density reachability are identified as noise points; therefore, this cluster is identified as the same cluster. During the movement of the weeding machine, each time new point cloud data is collected, the new data is merged with the historical point cloud data into a full dataset, and the DBSCAN clustering process described above is re-executed to repeatedly calculate the neighborhood, core points and cluster relationships for the full dataset. Based on the clustering results and the distribution pattern of rice, densely distributed clusters were identified as rice, and sparsely distributed or scattered small clusters were identified as weeds, thus obtaining the three-dimensional coordinate positions of rice and weeds. Step 4: During the forward movement of the weeder, when the pressure sensor (4) installed at the front end of the flexible rod (3) comes into contact with the rice seedling and is deformed by the pressure of the rice plant, it feeds back the pressure signal to the microcontroller inside the guide device, which calculates the difference and determines whether the weeder has deviated from the center of the rice row, and then controls the wheels to correct the path. Meanwhile, the integrated computer (5) transmits the rice and weed location data obtained in step 3 to its internal path planning module and weeding control module; the path planning module combines the position of the rice row centerline sensed by the pressure sensor (4) to plan the weeder's travel path, ensuring that it avoids the rice cluster coverage area and covers the weed cluster distribution area; the weeding control module controls the weeding wheel (16) between rows to rotate continuously as the weeder moves forward, and the curved design at the end of the blade helps to break up the soil and remove weeds between rows; When the clustering results show that there are weed clusters between plants, the integrated computer (5) sends a command to the motor C to drive the gear A (13) to rotate. By controlling the direction and angle of the gear A (13), the rotation of the driven gear B (24) is controlled. The driven gear B (24) drives the rotating shaft (18) to swing, which in turn drives the weeding wheels (17) between plants to rotate closer to each other, cutting the weed clusters between plants. The rice passes through the gap between the two weeding wheels to avoid damaging the seedlings.

Images