Weed control device and method based on autonomous weed identification

Abstract

This invention provides a weeding device and method based on autonomous weed identification, belonging to the field of intelligent agriculture technology. The device includes: a robot unit for automatically moving within a weeding area; a data collection unit mounted on the robot unit for collecting scene information within the visible area during movement; an identification unit for identifying weeds based on the scene information and locating weeds based on the identification results; a scheme generation unit for calculating blue laser dosage based on weed location and generating a corresponding weeding scheme based on the calculation results; and a weeding unit including a blue laser for executing the weeding scheme. This invention solves the problems of short battery life and low weeding efficiency of existing weeding robots.

Description

Technical Field

[0001] This invention relates to the field of intelligent agriculture technology, specifically to a weeding device and a weeding method based on autonomous weed identification. Background Technology

[0002] Grassland industry is an emerging industry that has emerged from traditional agriculture with the progress of society and economic development. It plays a key role in agricultural modernization and ecological civilization construction. Weeds pose a serious threat to the growth of pasture. Their rapid growth not only occupies the growth space of pasture and reduces the sunlight of crops, but also breeds pests, which directly affects the yield of pasture and may even lead to no harvest. Therefore, timely control of weeds in the field is crucial. At present, common weeding methods include manual weeding and chemical weeding[2]. Although manual weeding is highly accurate, it is labor-intensive, inefficient and costly. Although chemical weeding is efficient and economical, its extensive use may lead to waste of drugs and damage to similar pastures. With the improvement of agricultural mechanization, mechanical weeding methods have emerged. Its advantages are high efficiency and low manpower consumption, but large machinery has poor adaptability and may damage crops.

[0003] Utilizing advanced technologies for intelligent weed control offers a pollution-free and highly efficient way to remove weeds from fields. Therefore, researching intelligent weeding robots is a crucial way for the nation to improve its agricultural level. With technological advancements, the accuracy of weed identification is increasing, making fully automated weeding based on autonomous weed recognition increasingly a reality. However, weeds are diverse, and existing weeding robots often operate according to a uniform standard, unable to adapt to different weed types with varying power outputs, resulting in excessive energy consumption and short operating times. Furthermore, existing weeding robots are often unsuitable for complex weeding scenarios, and their flexibility cannot meet the demands for high-efficiency weed control. To address the problems of short operating times and low weeding efficiency in existing solutions, a new weeding device suitable for autonomous weed identification is needed. Summary of the Invention

[0004] The purpose of this invention is to provide a weeding device and method based on autonomous weed identification, so as to at least solve the problems of short battery life and low weeding efficiency of existing weeding robots.

[0005] To achieve the above objectives, the first aspect of the present invention provides a weeding device based on autonomous weed identification. The device includes: a robot unit for automatically moving within a weeding area; a data collection unit disposed on the robot unit for collecting scene information within a visible area during the movement; an identification unit for identifying weeds based on the scene information and locating weeds based on the identification results; a scheme generation unit for calculating blue laser dosage based on weed location and generating a corresponding weeding scheme based on the calculation results; and a weeding unit including a blue laser for executing the weeding scheme.

[0006] Optionally, the robot unit includes: a mobile chassis, built based on Mecanum wheels, for performing movement; an energy storage module for powering each unit; and multiple fixed modules disposed on the robot surface, respectively for mounting the acquisition unit and the weeding unit.

[0007] Optionally, the acquisition unit includes: a radar module for sensing obstacles on the movement path during the robot unit's movement; a positioning module for acquiring the real-time position information of the robot unit; and an image acquisition module for acquiring scene image information within the visible area during the movement.

[0008] Optionally, the step of calculating the blue laser dose based on weed location and generating a corresponding weeding plan based on the calculation results includes: determining the weed type based on the weed location results; and determining the linkage curve between the dry weight of weeds and the blue laser dose for the corresponding weed type in a pre-built knowledge base, represented as:

[0009]

[0010] Where DW is the dry weight of the weeds after laser treatment; Dose is the laser dose; C is the minimum dry weight of the weeds after laser treatment; and D is the maximum dry weight of the weeds without laser treatment. The laser dose and the position and tilt angle of the dry weight curve are used to describe the weed dry weight when it is reduced to 50% of DD; σ is an error amount of approximately zero; based on the linkage curve between the weed dry weight and the blue laser dose, the blue laser dose that meets the weed control requirements is determined, and a weed control plan is generated based on the calculation results.

[0011] Optionally, the weeding unit includes: a robotic arm and a blue laser disposed at the end of the robotic arm; the robotic arm is a five-degree-of-freedom robotic arm; the blue laser is always perpendicular to the ground, and the distance between the moving plane and the ground remains unchanged.

[0012] Optionally, the robotic arm includes two rotary kinematic pairs and three swing kinematic pairs; the robotic arm is provided with a first rotary kinematic pair, a first swing kinematic pair, a second rotary kinematic pair, a second swing kinematic pair, and a third swing kinematic pair in sequence from the end equipped with the blue laser to the root of the robotic arm; the kinematic pairs are connected to each other based on the arm lever.

[0013] Optionally, the robotic arm includes four arms, which are sequentially arranged from the end equipped with the blue laser to the root of the robotic arm as a first arm, a second arm, a third arm, and a fourth arm; the fourth arm is connected to the robotic arm turntable of the robot unit; the blue laser is located at the end of the first arm; the first arm and the second arm, the second arm and the third arm, and the third arm and the fourth arm are all connected by hinges, and a servo motor is provided to control their rotation; the fourth arm is connected to the robotic arm turntable by hinges, and two servo motors are provided to control their rotation.

[0014] A second aspect of the present invention provides a weeding method based on autonomous weed identification, the method comprising: automatically circulating within a weeding area and collecting scene information within the visible area during the circulation process; identifying weeds based on the scene information and locating weeds based on the identification results; calculating blue laser dose based on the weed location and generating a corresponding weeding plan based on the calculation results; and executing the weeding plan to generate blue laser light.

[0015] Optionally, the step of calculating the blue laser dose based on weed location and generating a corresponding weeding plan based on the calculation results includes: determining the weed type based on the weed location results; and determining the linkage curve between the dry weight of weeds and the blue laser dose for the corresponding weed type in a pre-built knowledge base, represented as:

[0016]

[0017] Where DW is the dry weight of the weeds after laser treatment; Dose is the laser dose; C is the minimum dry weight of the weeds after laser treatment; and D is the maximum dry weight of the weeds without laser treatment. The laser dose and the position and tilt angle of the dry weight curve are used to describe the weed dry weight when it is reduced to 50% of DD; σ is an error amount of approximately zero; based on the linkage curve between the weed dry weight and the blue laser dose, the blue laser dose that meets the weed control requirements is determined, and a weed control plan is generated based on the calculation results.

[0018] On the other hand, the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described weed removal method based on autonomous weed identification.

[0019] Through the above technical solutions, the present invention automatically transfers the weeding area to achieve automated operation; during the transfer process, it collects scene information within the visible area to improve data collection efficiency; it identifies weeds based on the scene information to achieve intelligent identification; it locates weeds accurately based on the identification results; it calculates the blue laser dose to scientifically determine the laser dose, effectively reducing losses and ensuring battery life; it generates corresponding weeding plans to provide personalized weeding solutions; it executes the weeding plans to achieve automated weeding operation; and it generates blue laser light to effectively implement the weeding plan.

[0020] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0021] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings:

[0022] Figure 1 This is a schematic diagram of the structure of a weeding device based on autonomous weed identification provided in one embodiment of the present invention;

[0023] Figure 2 This is a schematic diagram of the structure of a robotic arm provided in one embodiment of the present invention;

[0024] Figure 3 This is a flowchart of the steps of a weed removal method based on autonomous weed identification provided in one embodiment of the present invention.

[0025] Explanation of reference numerals in the attached figures

[0026] 01-First rotary kinematic pair; 02-First oscillating kinematic pair; 03-Second rotary kinematic pair; 04-Second oscillating kinematic pair; 05-Third oscillating kinematic pair; 06-First lever; 07-Second lever; 08-Third lever; 09-Fourth lever; 10-Robot arm turntable. Detailed Implementation

[0027] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0028] Figure 1 This is a structural diagram of a weeding device based on autonomous weed identification provided in one embodiment of the present invention. Figure 1As shown, this invention provides a weeding device based on autonomous weed identification. The device includes: a robot unit for automatically moving within a weeding area; a data collection unit disposed on the robot unit for collecting scene information within the visible area during the movement; an identification unit for identifying weeds based on the scene information and locating weeds based on the identification results; a scheme generation unit for calculating blue laser dosage based on weed location and generating a corresponding weeding scheme based on the calculation results; and a weeding unit including a blue laser for executing the weeding scheme.

[0029] In this embodiment of the invention, a robotic unit is used to automatically move the weeding area, improving weeding efficiency and reducing labor costs; a data collection unit collects scene information within the visible area during the movement, providing data support for subsequent weed identification and location; an identification unit identifies and locates weeds, accurately determining their positions; a scheme generation unit uses blue laser dose calculation technology to provide a scientific basis for formulating weeding schemes; and a weeding unit includes a blue laser, which uses advanced laser technology to execute the weeding scheme, achieving highly efficient weeding results.

[0030] Preferably, the robot unit includes: a mobile chassis, built based on Mecanum wheels, for performing movement; an energy storage module for powering each unit; and multiple fixed modules disposed on the robot surface, respectively for mounting the acquisition unit and the weeding unit.

[0031] In this embodiment of the invention, the Mecanum wheel operates on the principle of force analysis. When the Mecanum wheel rotates forward, the rollers on its rim generate a frictional force opposite to the direction of the wheel, thus propelling the vehicle forward. Similarly, when the Mecanum wheel rotates backward, the frictional force generated by the rollers causes the vehicle to move backward. By controlling the rotational speed and direction of each Mecanum wheel, a resultant force vector can be synthesized, enabling the vehicle to translate in any direction. The advantages of the Mecanum wheel are its high maneuverability and flexibility. It allows for precise positioning and control in confined spaces, making it suitable for applications requiring frequent changes in direction and position.

[0032] Based on this, Mecanum wheels can achieve lateral movement, meaning that robots or vehicles can move through narrow field paths or between crops without complex steering operations; because Mecanum wheels can translate, rotate, and move diagonally, robots or vehicles can more flexibly traverse field crops, avoid obstacles, and quickly adjust their direction; the precise control capabilities of Mecanum wheels enable robots or vehicles to achieve accurate positioning and movement in field scenarios, which is helpful for weeding operations; and because Mecanum wheels can move smoothly and operate in narrow spaces, they can reduce damage to surrounding crops and improve operational efficiency.

[0033] Preferably, the energy storage module includes a battery pack, which is installed at the rear of the device to provide power to various parts of the robot and also to balance the overall weight distribution of the robot.

[0034] Preferably, the acquisition unit includes: a radar module for sensing obstacles on the movement path during the robot unit's movement; a positioning module for acquiring the robot unit's real-time position information; and an image acquisition module for acquiring scene image information within the visible area during the movement.

[0035] In this embodiment of the invention, the radar module can emit radio waves and receive their reflections. By analyzing the intensity and timing of the reflected signals, it detects surrounding obstacles, such as trees, rocks, or other obstructions. This helps the robot avoid collisions and maintain a safe distance. It also helps the robot perceive the height and shape of the ground to adjust its movement path and avoid uneven terrain, which is crucial for moving and working in different types of field terrain. Furthermore, it can be used to locate the robot's position and navigate to designated target points. By combining it with data from other sensors (such as GPS and inertial navigation systems), radar can provide more accurate positioning information, ensuring the robot moves along a predetermined path. Radar can also be used to control the working range of the weeding robot, ensuring coverage of the area requiring weeding without exceeding the designated area. By detecting surrounding boundaries and obstacles, radar helps the robot work in the correct area.

[0036] Furthermore, to ensure accurate positioning of weed targets, the image acquisition module of this invention is based on a binocular camera. A binocular camera can simultaneously capture the scene using two cameras, thereby achieving depth perception and 3D reconstruction. By calculating the disparity between the two cameras, the distance and depth information of the weeds can be accurately estimated, which is crucial for accurate distance measurement and environmental perception. Binocular cameras are generally more robust than monocular cameras in varying lighting conditions and complex environments. By utilizing information from two cameras, the impact of lighting, shadows, or occlusion on a single camera can be reduced, improving the stability and reliability of the system. This invention's binocular camera-based scene image acquisition provides more comprehensive images for subsequent weed identification, ensuring higher accuracy.

[0037] Preferably, the recognition unit is built using the Jetson Nano development board, a compact yet powerful embedded AI development board. It boasts strong computing and parallel processing capabilities, making it suitable for deep learning, machine vision, and AI applications. It can handle complex algorithms and models, enabling rapid inference and computation. The Jetson Nano has relatively low power consumption, making it suitable for embedded systems and edge computing applications. It provides high-performance computing capabilities with limited power consumption, saving energy and extending device lifespan. Building with the Jetson Nano development board also reduces robot energy consumption, further improving robot endurance. The Jetson Nano development board supports a range of popular AI frameworks and algorithms, such as TensorFlow, PyTorch, Caffe / Caffe2, Keras, and MXNet, allowing developers to easily and quickly integrate AI models and frameworks into products, easily implementing powerful functions such as image recognition, object detection, pose estimation, semantic segmentation, video enhancement, and intelligent analysis, meeting the needs of weed target recognition.

[0038] Currently, many training models have emerged for weed target recognition. We will not elaborate on weed location recognition and weed type recognition. The solution of this invention only provides the corresponding weeding action based on the recognition result.

[0039] Preferably, the present invention utilizes blue laser for weed removal. Previous laser weeding robots mostly used CO2 laser emitters, which are infrared lasers with wavelengths typically around 10.6 micrometers. CO2 lasers have high energy density and strong penetrating power, capable of generating a thermal effect deep within weed tissue. However, CO2 laser energy absorption is mainly concentrated on the surface layer of weeds, potentially requiring longer irradiation times to remove thicker weed stems. Furthermore, CO2 laser devices are large and expensive, making them unsuitable for mobile weeding robots. In contrast, blue lasers have shorter wavelengths, typically between 400-500 nanometers. Blue lasers have high energy density and strong absorption capacity, generating a thermal effect on shallower weed surfaces. The energy absorption of blue lasers is mainly concentrated on the surface layer of weeds, causing faster burning and scorching. Moreover, blue laser devices are relatively small and inexpensive, making them more suitable for integration and application in mobile weeding robots.

[0040] Furthermore, blue laser was chosen as the light source for weed control primarily due to its advantages such as high energy density, strong absorption capacity, good combustion effect, and compact and economical equipment. While CO2 laser has a better weed-control effect at the same dosage compared to blue laser, CO2 laser emitters are larger and more expensive, require water-cooling radiators, and have stringent environmental requirements, making them unsuitable for small weed-control robots. Blue laser refers to a special type of laser with a wavelength between 360nm and 480nm and a blue light source. Blue lasers have characteristics such as short wavelength and low diffraction, showing promising application prospects. In practical applications, compared to other commonly used lasers, blue lasers are relatively inexpensive and have advantages such as small size, stable operation, and long lifespan. Currently, blue lasers are widely used in printing, laser cutting, screen display, biochemistry, and optical information storage. The blue laser emitter selected for the weed-control robot uses a 12V DC power supply.

[0041] In one possible implementation, during actual weeding operations, a robotic arm controls the movement direction of a blue laser, which cuts the stems of the weeds. The laser spot on the ground can be considered a point; the laser intensity at this point depends on the laser power and the speed at which the spot moves. This laser intensity is also called the laser Dosw (J / mm) dose, expressed as:

[0042]

[0043] Where P represents laser power in W, and V represents the laser beam velocity in mm / s. For the same laser power, a faster moving speed results in a smaller laser dose, while a slower moving speed results in a larger laser dose. Selecting an appropriate laser dose is crucial to this design. This not only enables precise weed removal but also saves energy and increases the robot's operating time.

[0044] Preferably, the step of calculating the blue laser dose based on weed location and generating a corresponding weeding plan based on the calculation results includes: determining the weed type based on the weed location results; and determining the linkage curve between the dry weight of weeds and the blue laser dose for the corresponding weed type in a pre-built knowledge base, represented as:

[0045]

[0046] Where DW is the dry weight of the weeds after laser treatment; Dose is the laser dose; C is the minimum dry weight of the weeds after laser treatment; and D is the maximum dry weight of the weeds without laser treatment. The laser dose and the position and tilt angle of the dry weight curve are used to describe the weed dry weight when it is reduced to 50% of DD; σ is an error amount of approximately zero; based on the linkage curve between the weed dry weight and the blue laser dose, the blue laser dose that meets the weed control requirements is determined, and a weed control plan is generated based on the calculation results.

[0047] In this embodiment of the invention, the plant's response to laser is influenced by various factors, including plant species, tissue structure, leaf thickness, and pigment content. Therefore, even with the same laser dose, different plants may exhibit different removal effects. Some plants may be more sensitive to lasers and easier to remove, while others may require higher laser doses to achieve the same removal effect. The physiological characteristics and tissue structure of plants affect the absorption and propagation of lasers within plant tissues, thus influencing the laser's effectiveness. Therefore, when using lasers for plant removal or treatment, it is necessary to adjust the laser dose and parameters according to the specific plant type and purpose to achieve the desired effect. In practice, experiments and adjustments may be required to determine the most suitable laser treatment parameters for a particular plant.

[0048] Therefore, if the same laser dose is used to remove all weeds, to ensure sufficient removal effect, the laser dose produced each time must meet the laser dose requirement of the weed with the highest dose demand. This requires the laser emitter to always maintain a high operating power, resulting in a shortened runtime. Therefore, the solution of this invention adjusts the laser dose differently based on the weed type, which can ensure weed removal effect while reducing system losses and extending runtime.

[0049] In this embodiment of the invention, the research by Streibig et al. shows that after irradiation with gradually increasing laser doses, the trend of the plant's dry weight (DW) is an S-shaped decreasing curve, where the minimum dry weight (C) is approximately 0. When the laser dose (D) approaches infinity, the weed dry weight (D) is approximately equal to the minimum dry weight (C), indicating that high-dose laser irradiation kills the plant, causing its dry weight to approach 0. When no laser irradiation is used and the laser dose (Dose) is 0, the weed dry weight (D) is approximately equal to the maximum dry weight (D), indicating that plants not irradiated by lasers grow normally.

[0050] Preferably, to more intuitively describe the weed-control effect of blue laser and adapt to the specific requirements of field weed control, referring to the standards for evaluating herbicides, the concept of dry weight efficacy is introduced to describe the weed-control effect after laser treatment, expressed as:

[0051]

[0052] Where Y represents the dry weight control efficacy; Ck represents the dry weight of weeds in the un-laser-irradiated control area (in g); and E represents the dry weight of weeds in the laser-irradiated test area (in g). Weeds at the same growth stage were cultivated indoors under constant growing conditions and divided into test and control areas. Laser irradiation was conducted, and after a certain period, the difference between the dry weight of weeds in the test area and the dry weight of weeds in the un-laser-irradiated control area was calculated. The percentage of this difference relative to the dry weight of weeds in the control area was defined as the dry weight control efficacy of blue laser on weeds at that growth stage. According to the research of Streibig et al., within a certain laser dose range, the dry weight control efficacy Y of weeds should also exhibit an S-shaped curve. In the determination of herbicide control of various weeds, a dry weight control efficacy of over 80% is generally considered a good control effect. To obtain good weed control effects while controlling the laser dose, accelerating the weed control operation, and avoiding energy waste, the target dry weight control efficacy Y was set at 87%, referring to the control standards for herbicides.

[0053] Preferably, the weeding unit includes: a robotic arm and a blue laser disposed at the end of the robotic arm; the robotic arm is a five-degree-of-freedom robotic arm; the blue laser is always perpendicular to the ground, and the distance between the moving plane and the ground remains unchanged.

[0054] In this invention, the robotic arm is the most widely used automated mechanical device in the field of robotics. Generally speaking, controlling the end effector of the robotic arm to move to any point in three-dimensional space requires six degrees of freedom, namely three rotational kinematic pairs and three oscillating kinematic pairs.

[0055] The formula for calculating the degree of freedom F of a spatial mechanism is:

[0056] F = 6n - 5P5 - 4P4 - 3P3 - 2P2 - P1

[0057] In the formula, n represents the number of moving components, p5 represents a component with 1 degree of freedom and 5 constraints, and so on. The weeding robot designed in this invention uses a robotic arm to adjust the position of the laser emitter. During weeding operations, the laser emitter remains perpendicular to the ground and moves on a plane at a certain height above the ground. In other words, the robotic arm of the weeding robot designed in this invention only needs to control the laser emitter to move flexibly on a fixed plane, making the design requirements relatively simple. Therefore, a revolute joint was removed from the end of the robotic arm, reducing one degree of freedom, resulting in a five-degree-of-freedom robotic arm structure.

[0058] Preferred, such as Figure 2The robotic arm includes two rotary kinematic pairs and three swing kinematic pairs; the robotic arm is provided with a first rotary kinematic pair 01, a first swing kinematic pair 02, a second rotary kinematic pair 03, a second swing kinematic pair 04, and a third swing kinematic pair 05 in sequence from the end equipped with the blue laser to the root of the robotic arm; the kinematic pairs are connected by the arm rod.

[0059] Furthermore, the robotic arm includes four arms, which are, sequentially from the end equipped with the blue laser to the root of the robotic arm, a first arm 06, a second arm 07, a third arm 08, and a fourth arm 09; the fourth arm 09 is connected to the robotic arm turntable 10 of the robot unit; the blue laser is located at the end of the first arm 06; the first arm 06 and the second arm 07, the second arm 07 and the third arm 08, and the third arm 08 and the fourth arm 09 are all connected by hinges, and a servo motor is provided to control their rotation; the fourth arm 09 is connected to the robotic arm turntable 10 by hinges, and two servo motors are provided to control its rotation.

[0060] In this embodiment of the invention, since the servo motors at the fourth arm 09 and the robotic arm turntable 10 need to control the lifting and lowering of the entire robotic arm and have the largest load, two servo motors are used to drive them simultaneously to improve the overall stability of the robotic arm during movement. The robotic arm turntable 10 is hinged to the frame of the whole machine and its relative rotation with the frame is controlled by a servo motor. The robotic arm turntable 10 is responsible for driving the entire robotic arm to rotate.

[0061] Preferably, executing the weeding scheme includes: determining the movement rules of the robotic arm based on the weed location results and a pre-constructed robotic arm movement model; wherein the robotic arm movement model is constructed based on the DH method; generating control commands for each servo motor based on the movement rules of the robotic arm; and generating blue laser light of the corresponding laser dose based on the weeding scheme after the robotic arm movement is completed.

[0062] Specifically, in the process of modeling the robotic arm, the DH method is used to analyze the four-axis robotic arm, which can better characterize the structure of the robotic arm. The DH method is a method for analyzing the structure of robotic arms proposed by Denavit and Hartenberg. This method fixes a coordinate system on each link of the robot; then uses a 4x4 homogeneous transformation matrix to describe the spatial positional relationship between two links, and finally derives the equivalent homogeneous transformation matrix of the robotic arm's "end-effector coordinate system" relative to the "reference frame", and establishes the motion equation of the manipulator.

[0063] In one possible implementation, the links are numbered from 0 to n, where link 0 generally represents the robot arm base and link n generally represents the end effector. The i-th joint connects the (i-1)-th link and the i-th link. In this way, a robot arm with n degrees of freedom contains a total of n+1 links and n joints. Links are connected via R-joints or P-joints. In the DH method, it's important to note that the coordinate system fixed to the (i-1)-th link is Fi, and the coordinate system fixed to the i-th link is Fi+1. The origin of coordinate system Fi is defined as Oi, and the axes are Xi, Yi, and Zi. Then, in the DH method, the rules for establishing the first n coordinate systems are as follows:

[0064] 1) Zi represents the axis of the i-th joint, but the direction of Zi is not defined. Therefore, in practical applications, we may see this axis with two positive and negative directions. Especially under special conditions, such as the Zi axis of a prism joint can be placed in any position, because the prism joint only has a defined direction (not a specific position).

[0065] 2) Unlike Zi, Xi's positive direction is defined from the outset. Xi represents a line perpendicular to both the Zi-1 axis and the Zi axis, with its direction pointing from Zi-1 to Zi. There are three cases: in the case of inclination, the Xi axis is easily identified; in the case of intersection, the positive and negative directions of the Xi axis cannot be defined using the above method; therefore, we stipulate that the right-hand rule is used to define the positive direction of the Xi axis in this case. That is, if we use ii, Ki-1, and Ki to represent the unit vectors of Xi, Zi-1, and Z axes respectively, ii is defined as Ki-1Ki; in the case of parallelism, the position of the X-axis is determined by choosing a point through which the Xi axis passes to the (i-1)th coordinate system origin.

[0066] Figure 3 This is a flowchart of a weeding method based on autonomous weed identification provided in one embodiment of the present invention. Figure 3 As shown, this invention provides a weeding method based on autonomous weed identification, the method comprising:

[0067] Step S10: Automatically move within the weeding area and collect scene information within the visible area during the movement.

[0068] Step S20: Identify weeds based on the scene information and locate weeds based on the identification results.

[0069] Step S30: Calculate the blue laser dose based on weed location, and generate the corresponding weed control plan based on the calculation results.

[0070] Step S40: Execute the weeding scheme to generate blue laser light.

[0071] The present invention also provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the aforementioned xx.

[0072] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a microcontroller, chip, or processor to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0073] The optional embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details described above. Within the scope of the technical concept of the embodiments of the present invention, various simple modifications can be made to the technical solutions of the embodiments of the present invention, and these simple modifications all fall within the protection scope of the embodiments of the present invention. It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the embodiments of the present invention will not further describe the various possible combinations.

[0074] Furthermore, various different embodiments of the present invention can be combined in any way, as long as they do not violate the spirit of the embodiments of the present invention, they should also be regarded as the content disclosed by the embodiments of the present invention.

Claims

1. A weeding device based on autonomous weed identification, characterized in that, The device includes: Robotic units are used to automatically move around in the weeding area; A data acquisition unit, mounted on the robot unit, is used to acquire scene information within the visible area during the workflow; wherein, The acquisition unit includes: a radar module for sensing obstacles on the movement path of the robot unit during its movement; a positioning module for acquiring the real-time position information of the robot unit; and an image acquisition module for acquiring scene image information within the visible area during the movement. The identification unit is used to identify weeds based on the scene information and locate weeds based on the identification results; The scheme generation unit is used to calculate the blue laser dose based on weed location and generate a corresponding weeding scheme based on the calculation results; among which... The calculation of blue laser dose based on weed location, and the generation of corresponding weed control plans based on the calculation results, include: determining the weed type based on the weed location results; and determining the linkage curve between weed dry weight and blue laser dose for the corresponding weed type in a pre-built knowledge base, represented as follows: Where W is the dry weight of the weeds after laser treatment; ose is the laser dose; t is the minimum dry weight of the weeds after laser treatment; and t is the maximum dry weight of the weeds without laser treatment. The position and tilt angle of the laser dose and dry weight curve are used to describe the position of the curve when the dry weight of weeds is reduced to 50% of 𝐷𝐷; 𝜎 is an error amount of approximately zero; based on the linkage curve between the dry weight of weeds and the blue laser dose, the blue laser dose that meets the weed control requirements for the current weeds is determined, and a weed control plan is generated based on the calculation results. The weeding unit includes a blue laser for executing the weeding scheme; wherein, The robot unit includes: a mobile chassis, built on Mecanum wheels, for moving; an energy storage module for powering each unit; and multiple fixed modules disposed on the robot surface, for mounting the acquisition unit and the weeding unit, respectively.

2. The apparatus according to claim 1, characterized in that, The weeding unit includes: A robotic arm and a blue laser located at the end of the robotic arm; The robotic arm is a five-degree-of-freedom robotic arm; The blue laser is always perpendicular to the ground, and the distance between the moving plane and the ground remains unchanged.

3. The apparatus according to claim 2, characterized in that, The robotic arm includes two rotary kinematic pairs and three oscillating kinematic pairs; The robotic arm is provided with a first rotary kinematic pair, a first swing kinematic pair, a second rotary kinematic pair, a second swing kinematic pair, and a third swing kinematic pair in sequence from the end equipped with the blue laser to the root of the robotic arm; The kinematic pairs are connected by lever arms.

4. The apparatus according to claim 3, characterized in that, The robotic arm includes four arms, which are the first arm, the second arm, the third arm, and the fourth arm, arranged sequentially from the end equipped with the blue laser to the root of the robotic arm. The fourth arm is connected to the robotic arm turntable of the robot unit; The blue laser is located at the end of the first arm; The first arm and the second arm, the second arm and the third arm, and the third arm and the fourth arm are all connected by hinges, and a servo motor is used to control the rotation. The fourth arm is connected to the robotic arm turntable by a hinge, and two servo motors are used to control its rotation.

5. The apparatus according to claim 4, characterized in that, The execution of the weeding plan includes: Based on the weed location results and a pre-constructed robotic arm movement model, the movement rules of the robotic arm are determined; among them, The robotic arm movement model is constructed based on the DH method; Control commands for each servo motor are generated based on the movement rules of the robotic arm, and after the robotic arm completes its movement, a corresponding dose of blue laser light is generated based on the weeding scheme.

6. A weeding method based on autonomous weed identification, characterized in that, The method is implemented based on the weed removal device based on autonomous weed identification as described in any one of claims 1-5, and the method includes: It automatically moves around the weeding area and collects scene information within the visible area during the movement. Weeds are identified based on the scene information, and their location is determined based on the identification results. Blue laser dose is calculated based on weed location, and corresponding weed control plan is generated based on the calculation results. The weeding process is executed to generate a blue laser.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the weed removal method based on autonomous weed identification as described in claim 6.

Images