Intelligent pruning method and system for green plants, pruning robot and control device

Abstract

The invention relates to the field of landscaping maintenance, and discloses an intelligent pruning method and system for green plants, a pruning robot and a control device.The method comprises the following steps that real-time data of the plants are obtained through a sensor; evaluating data by using machine learning and making a pruning plan; generating an optimal travel path based on the plan; controlling the robot to perform trimming according to the path and feed back force and angle; adjusting a pruning strategy according to the feedback and outputting a result; after trimming is completed, evaluating a result and adjusting a path or a strategy; the invention further provides an intelligent pruning system for the green plants. A machine learning module; a path planning module; a control module; a feedback module; and an evaluation module. The intelligent analysis technology based on multi-sensor data acquisition and machine learning is adopted, the growth state of the plant can be obtained in real time, the pruning plan can be dynamically generated, and the efficient and accurate pruning effect is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of garden greening maintenance, specifically to a method, system, pruning robot and control device for intelligent pruning of green plants. Background Art

[0002] In the existing technology, pruning tasks in urban park green spaces, residential area landscape greening, and green belts on both sides of highways are mostly completed manually. Manual pruning is not only labor-intensive but also inefficient, especially in large-scale greening projects, where labor costs are extremely high. Due to reliance on manual judgment, the effect of pruning is often limited by the experience and skills of the operator, making it difficult to achieve consistency and accuracy. Moreover, manual pruning has limited ability to judge the growth status of plants, and it is often difficult to discover and respond to the different needs of plants in real time, resulting in over-pruning or under-pruning during the pruning process, which ultimately affects the health of the plants and the landscape effect.

[0003] At the same time, although existing automatic pruning equipment has been used in certain scenarios, most of them use fixed path planning methods for pruning. Such solutions lack intelligent recognition and analysis of plant status, and it is difficult to make dynamic adjustments based on the growth conditions of the plants. Pruning robots rely on preset paths to operate. In environments with complex or diverse plant forms, it is difficult to accurately avoid obstacles or fail to fully identify the growth characteristics of plants, which in turn affects the pruning effect. In addition, existing automatic pruning equipment generally lacks sufficient flexibility, making it difficult to adjust pruning strategies in a timely manner based on the plant's response during the pruning process, and it is difficult to ensure that every branch and leaf is properly pruned.

[0004] While some pruning equipment includes basic sensor functionality, these sensors' operating range and analytical capabilities are typically limited. Existing sensors primarily focus on simple obstacle avoidance, failing to comprehensively monitor and analyze the specific needs of plants. Existing equipment struggles to provide real-time assessments of plant health, neglecting key physiological characteristics like moisture and chlorophyll content, making it difficult to effectively prevent plant health issues.

[0005] Therefore, the present invention proposes a method, system, pruning robot and control device for intelligent pruning of green plants to solve the shortcomings of the existing technology. Summary of the Invention

[0006] In response to the shortcomings of the existing technology, the present invention provides an intelligent pruning method, system, pruning robot and control device for green plants, which solves the problem that the automatic pruning equipment in the existing technology is difficult to dynamically adjust the pruning strategy according to the growth status of the plants, resulting in inconsistent pruning effects and low efficiency.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: The green plant intelligent pruning method includes the following steps: Acquire real-time data of plants in the area to be pruned through a variety of sensors; Use machine learning models to evaluate real-time plant data to develop personalized pruning plans; generating an optimal travel path for the pruning robot based on the pruning plan; Controlling the pruning robot to execute the pruning plan according to the optimal travel path and providing real-time feedback information, wherein the real-time feedback information includes at least the force and angle of the pruning tool; During the pruning process, the pruning strategy is adjusted according to the real-time feedback information and the pruning results are output; After the pruning task is completed, the pruning results are evaluated in real time. If deviations are found, the optimal travel path or strategy of the pruning robot is adjusted to optimize the effect.

[0008] The present invention also provides a green plant intelligent pruning system, comprising: A sensor module for collecting real-time data of the plants to be pruned; Machine learning module for developing pruning plans based on real-time data; The path planning module is used to generate the optimal travel path of the pruning robot according to the pruning plan and adjust the path to meet the requirements of the pruning task; A control module, used for controlling the pruning robot to execute a pruning plan according to an optimal travel path; The feedback module is used to monitor the execution of the pruning plan of the pruning robot in real time, collect the force and angle of the pruning tool and the response data of the plant, and output the generated pruning results; The evaluation module is used to evaluate the pruning results after the pruning task is completed and adjust the pruning path or strategy to optimize the pruning effect.

[0009] The present invention also provides an intelligent pruning robot for green plants, comprising: Vision sensors, lidar, near-infrared sensors, and force sensors to capture plant images, 3D morphology, health data, and the forces applied during pruning; The processing unit is used to formulate pruning plans based on real-time data and control the force, angle and path adjustment of the pruning tools.

[0010] The present invention also provides a green plant intelligent pruning control device, comprising: Data acquisition module, used to collect real-time data of plants and transmit it to the machine learning module; The evaluation module is used to formulate a pruning plan based on the data and coordinate the path planning module to control the execution of the robot; Feedback module, used to provide real-time status of pruning tools and plant response, and output generated pruning results; Pruning control module, used to adjust pruning tasks according to pruning results The present invention provides a method, system, pruning robot, and control device for intelligent pruning of green plants. The method has the following beneficial effects: 1. The method utilizes intelligent analysis technology based on multi-sensor data acquisition and machine learning, capable of acquiring plant growth status in real time and dynamically generating pruning plans, achieving efficient and accurate pruning results. Compared to existing solutions that rely on manual judgment and fixed pruning paths, the method can automatically adjust pruning strategies based on the actual plant conditions, avoiding over-pruning or under-pruning and significantly improving pruning accuracy and consistency.

[0011] 2. This invention utilizes a real-time feedback mechanism to monitor and adjust the plant's response, the force and angle of the pruning tool, and other factors during the pruning process, ensuring optimal pruning results. This compares to traditional pruning methods, which struggle with instant adjustment and feedback, often leading to unsatisfactory pruning results due to human error or operator error. This invention, through real-time feedback and dynamic adjustments, ensures the accuracy of each pruning operation and maximizes plant health.

[0012] 3. Through the collaborative work of path planning and pruning control modules, this invention achieves real-time optimization of the pruning robot's path and automatic adjustment of pruning tasks, thereby improving pruning efficiency and reducing labor intensity. Compared to existing traditional manual pruning methods, which are not only labor-intensive but also inefficient, this invention uses robots to automatically perform pruning, significantly reducing manpower requirements and improving efficiency, making it suitable for large-scale landscaping projects.

[0013] 4. This invention combines multiple advanced sensors, including visual sensors, lidar, and force sensors, to achieve comprehensive plant scanning and precise assessment, ensuring that every part of the plant is properly treated during the pruning process. Compared to existing automated pruning equipment, which lacks comprehensive plant identification and dynamic adjustment, this invention, through the integration of high-precision sensors, improves the consistency and aesthetics of pruning, avoiding the drawbacks of traditional pruning methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 is a flow chart of the method of the present invention; Figure 2 This is a system architecture diagram of the present invention; Figure 3 This is a schematic structural diagram of the pruning robot of the present invention; Figure 4 Schematic diagram of the control device of the present invention. DETAILED DESCRIPTION

[0015] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the present specification. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0016] See also Figure 1 The embodiment of the present invention provides a method for intelligent pruning of green plants, comprising the following steps: S1. Obtain real-time data of plants in the area to be pruned through various sensors; S2. Use machine learning models to evaluate real-time plant data and develop personalized pruning plans; S3. Based on the pruning plan, generating an optimal travel path for the pruning robot; S4, controlling the pruning robot to execute the pruning plan according to the optimal travel path and providing real-time feedback information, wherein the real-time feedback information includes at least the force and angle of the pruning tool; S5. During the pruning process, adjust the pruning strategy according to the real-time feedback information and output the pruning results; S6. After the pruning task is completed, the pruning results are evaluated in real time. If deviations are found, the optimal travel path or strategy of the pruning robot is adjusted to optimize the effect.

[0017] Step S1 uses multiple sensors to acquire real-time data on plants within the area to be pruned, providing essential data support for subsequent pruning tasks. This step provides foundational data for pruning plan development, path planning, pruning execution, and pruning result evaluation, and is a key step in implementing intelligent pruning and dynamically adjusting pruning strategies. Through this step, the system collects comprehensive and accurate plant data, providing a basis for optimizing subsequent modules.

[0018] In this embodiment, real-time data on plants within the area to be pruned is acquired using a variety of sensors. These sensors include visual sensors, lidar, near-infrared sensors, and force sensors. Each sensor has different functions and advantages, providing relevant plant data from different angles and dimensions.

[0019] Specifically, vision sensors are primarily used to capture plant morphological data, including images of leaves, branches, and other features. By analyzing these images, we can identify characteristics such as plant health, leaf distribution, and branch morphology. This morphological data is crucial for developing subsequent pruning strategies. For example, morphological data helps assess crown density and branch distribution, and helps machine learning models make accurate decisions during pruning.

[0020] Alternatively, lidar can be used to obtain three-dimensional plant morphological information. LiDAR technology emits a laser beam and receives the reflected signal, measuring the distance to each point on the plant surface to construct a three-dimensional image of the plant. LiDAR data not only helps determine the plant's geometric shape but also accurately captures parameters such as height, width, and depth. This information is critical for robot path planning and spatial obstacle avoidance during pruning.

[0021] In one possible implementation, near-infrared sensors are used to monitor plant health. Because a plant's health is closely linked to its physiological characteristics, near-infrared sensors can analyze physiological indicators such as moisture content and chlorophyll concentration through reflected light. For example, healthy plants typically exhibit high chlorophyll content, while plants with dry leaves or insufficient water will exhibit different reflectance characteristics. This data can help assess whether a plant is at the optimal time for pruning, thereby avoiding overpruning unhealthy plants.

[0022] Typically, data from multiple sensors is fused to improve accuracy and completeness. Data fusion technology can combine visual data, lidar data, near-infrared data, and other data to generate comprehensive plant data sets, further enhancing plant analysis capabilities. For example, combining image processing technology with point cloud data can provide higher-resolution 3D models, enabling more precise analysis of plant growth status.

[0023] Specifically, the fusion of image processing and 3D point cloud data can be achieved through image registration algorithms. These algorithms align and optimize data collected by different sensors, enabling precise matching of data from different viewpoints and devices. This process helps eliminate errors caused by factors such as perspective and resolution differences, thereby ensuring final data accuracy.

[0024] Alternatively, filtering algorithms can be used during data processing to remove noise and improve data quality and reliability. For example, Kalman filtering effectively addresses fluctuations and uncertainties in sensor data, resulting in more stable and reliable plant data. This high-quality data fully supports subsequent pruning schedules and path planning.

[0025] In some embodiments, before pruning begins, the system assesses the plant's health and growth needs based on acquired data. This assessment is based on real-time plant data. For example, if the near-infrared sensor indicates poor plant health, the system may automatically adjust the pruning schedule to avoid further damage or postpone pruning.

[0026] In some embodiments, the system can calculate the growth requirements and pruning targets of plants through formulas. Assume that the morphological data of the plant is D shape , health data is D health , the pruning target is T prune , can be evaluated by the following formula: T prune =f(D shape ,D health ); Here, f is a function obtained through machine learning training, which can generate a suitable pruning target based on the morphology and health status of the plant.

[0027] When planning the path, the robot needs to avoid certain parts of the plant to ensure the accuracy and safety of pruning. Assume that the robot's path is P robot , the obstacle area generated by the sensor data is A obstacle , the goal of path planning is P final , it can be calculated by the following formula: P final =P robot -A obstacle ; The final P calculated by this formula final All obstacle areas will be avoided to ensure that the robot can safely perform pruning tasks along the optimized path.

[0028] This embodiment achieves comprehensive acquisition of plant morphological characteristics, health status, and growth trends through the collaborative work of multiple sensors. This process not only provides accurate data input for machine learning models but also provides sufficient technical support for subsequent steps such as path planning, pruning plan formulation, pruning task execution, and result evaluation. Through efficient data fusion, image processing, point cloud data optimization, and filtering algorithms, the system's accuracy and stability can be improved, ensuring efficient and accurate pruning operations.

[0029] Step S2 develops a personalized pruning plan based on the real-time plant data acquired in step S1. By applying machine learning models, the system extracts valuable information from data such as plant morphology, health status, and growth trends, and uses this information to generate a pruning plan tailored to the plant's needs. This step provides detailed guidance for subsequent pruning tasks, ensuring that each plant is precisely pruned according to its individual growth profile.

[0030] In this embodiment, the core task of the machine learning model is to evaluate the real-time plant data collected by various sensors. The raw data collected by the sensors is preprocessed and then input into the machine learning algorithm for analysis. By learning the plant's morphological data, health data, and growth trend data, the model can automatically assess the plant's growth needs and generate a personalized pruning plan. It is worth noting that the processing and evaluation of the data is not limited to visual data, but also includes deep physiological data, such as the plant's moisture content, chlorophyll concentration, etc.

[0031] Specifically, machine learning models typically use deep learning algorithms, such as convolutional neural networks (CNNs) or recurrent neural networks (RNNs), which can learn plant growth patterns and regularities from a large amount of historical pruning data and automatically adjust pruning plans based on new real-time data. For example, the model can predict the future growth trend of the plant based on its current state, thereby providing the appropriate time for pruning. Through training data, the system can determine whether the plant needs to reduce branches, prune excessive leaves, or change the angle of certain branches to ensure healthy growth of the plant.

[0032] Optionally, machine learning models can incorporate historical plant data when assessing plant health. For example, by analyzing past pruning data and combining it with the plant's current growth status, the system can determine the optimal time to prune the plant. This assessment relies not only on the plant's appearance but also considers changes in its physiological characteristics, such as water loss or chlorophyll reduction, to develop a more informed pruning strategy.

[0033] In one possible implementation, a machine learning model assesses a plant's current pruning needs by comprehensively analyzing its morphological characteristics (such as crown density and branch angles) and health status (such as leaf color and photosynthesis level). The system automatically generates a pruning plan, including parameters such as pruning force, pruning angle, and pruning location. Each pruning task is formulated with optimal plant growth in mind to ensure that subsequent growth is not affected.

[0034] In some embodiments, to ensure personalized pruning plans, the system also makes targeted adjustments based on the characteristics of different plant types. For example, for ornamental plants, the pruning plan will consider maintaining their aesthetic appearance; for fruit trees, the pruning plan will optimize the pruning path based on the distribution of fruit to promote fruit growth and ripening.

[0035] Assume that the machine learning model estimates the optimal pruning target T through real-time plant data prune , and give pruning parameters according to growth requirements, including pruning force F, pruning angle θ and pruning position P location. Then the pruning objective can be expressed by the following formula: T prune =f(D shape ,D health ,D growth ); Among them, f is a trained deep learning model function; D shape It is the morphological data of the plant, including branch distribution, leaf density, etc.; D health It is the health status data of the plant, including water content, chlorophyll concentration, etc.; D growth It is the growth trend data of the plant, including historical pruning data, growth cycle, etc.

[0036] According to this formula, the model generates a suitable pruning solution T prune , which includes the trimming force F, trimming angle θ and trimming position P location , in order to achieve the best pruning effect.

[0037] Typically, the generated pruning plan is output as a multi-dimensional vector, encompassing all key factors in the plant pruning process. The machine learning model not only provides a pruning plan before each pruning session but also adjusts pruning tasks based on real-time data, optimizing the plan in real time.

[0038] Alternatively, the pruning plan can be adjusted through multiple iterations. For example, after each pruning cycle, the system automatically evaluates the pruning results and uses the model to adjust the subsequent pruning plan. This dynamic pruning strategy effectively avoids over-pruning or under-pruning, ensuring that the plant grows in the desired direction after pruning.

[0039] In practice, through continuous learning and optimization of machine learning models, the system can flexibly adjust to different plant species and growth environments. This makes the entire pruning process highly adaptable and intelligent, improving not only pruning accuracy but also efficiency and sustainability.

[0040] In this embodiment, step S2 analyzes real-time plant data using a machine learning model to develop a personalized pruning plan. This process combines data on plant morphology, health, and growth trends to ensure a scientific and targeted pruning plan. By continuously optimizing the pruning plan, the system can provide precise, personalized pruning solutions tailored to the unique needs of each plant, significantly improving the efficiency and quality of pruning.

[0041] Step S3 involves generating an optimal path for the pruning robot based on the generated pruning plan. This step is performed after the pruning plan is developed in step S2. The pruning robot must execute the actual operations according to the pruning plan, so path planning is crucial to ensuring the successful completion of the pruning task. By generating an optimal path, the pruning robot can efficiently and accurately move within the pruning area, avoid obstacles, and complete the task according to the set pruning plan.

[0042] In this embodiment, the process of generating the optimal path for the pruning robot based on the pruning plan first relies on the pruning plan and plant data generated in step S2. The pruning plan includes information such as specific pruning locations, pruning force, and angle, while also taking into account the characteristics of each plant and its position within the pruning area. Therefore, the path planning module needs to use this information to determine the robot's movement route, ensuring that the robot can successfully reach each target plant for pruning.

[0043] Specifically, the path planning module uses advanced algorithms, such as the A algorithm, Dijkstra's algorithm, or more complex deep learning-based path planning methods, to generate an optimal path based on the pruning plan and environmental constraints. The goal of path planning is to avoid collisions between the robot and other plants or obstacles through appropriate path planning, while also reducing pruning time and improving efficiency. For example, if the pruning task for a particular plant is complex, the path planning module will use the required pruning parameters to ensure that the robot reaches the correct location and prunes according to the pruning plan.

[0044] Alternatively, the path planning module not only calculates the path based on the target location provided by the pruning plan, but also takes into account dynamic changes in the environment in real time. For example, during the pruning process, changes in the surrounding vegetation or interference from the external environment may affect the accuracy of the path planning. To address these dynamic changes, the path planning module combines data from lidar and other sensors to dynamically adjust the robot's path to ensure the smooth and continuous execution of the pruning task.

[0045] In one possible implementation, the path planning module can also incorporate the robot's kinematic characteristics for optimization. Specifically, the robot's mobility, turning radius, and the working range of the pruning tool all influence the generation of an optimal path. By considering these kinematic parameters, the path planning module can calculate a path that aligns with the robot's motion characteristics, avoiding inefficient pruning operations caused by improper turns or insufficient range of motion.

[0046] In some embodiments, the path planning module may also incorporate a strategy based on the order in which plants are pruned. For example, for complex pruning tasks or large pruning areas, the path planning module must not only consider how the robot avoids obstacles but also prioritize pruning tasks to ensure the robot prunes each plant in the predetermined order to avoid duplication or missed pruning targets.

[0047] Typically, the pruning robot's speed is also considered during path planning. The path planning module dynamically adjusts the robot's speed based on the distribution of plants, the complexity of the pruning task, and the robot's operating speed. For simple pruning tasks, the robot can accelerate, while slowing down to ensure precise pruning in areas requiring finer detail.

[0048] Specifically, during path planning, the system also considers both static and dynamic obstacles in the environment. For example, the path planning module calculates the optimal path in real time based on obstacle data collected by the LiDAR sensor. The type and location of obstacles influence the path planning module's decisions, ensuring that the pruning robot avoids irrelevant objects such as trees, rocks, and equipment, thereby preventing collisions that could interrupt pruning or damage equipment.

[0049] Alternatively, the path planning module can evaluate the quality of a path using a predefined objective function. This objective function might include multiple optimization goals, such as shortest path, minimum time, and minimum energy consumption. In some cases, the system might adjust these goals based on the urgency of the pruning task to ensure that it can be completed efficiently within the specified timeframe.

[0050] In order to accurately generate the optimal path for the pruning robot, the path planning module calculates the shortest path and obstacle avoidance path based on the position data and environment data in the pruning task. Assume that the position of the pruning task is P target , the obstacle area is A obstacle , the robot's current position is P robot , the optimal path is P final , then the path planning goal can be expressed by the following formula: P final =argmin P [Distance(P,P target )+λ·Avoidance(P,A obstacle )]; Where P is any point on the path, representing each position of the robot as it moves along the path; Distance(P,P target ) is the current position P to the target position P targetThe Euclidean distance represents the shortest path for the robot to reach the target; Avoidance (P, A obstacle ) is the distance from the current position P to the obstacle area A obstacle The obstacle avoidance cost function represents the distance from the obstacle on the path. The farther away, the lower the cost, and the closer, the higher the cost. λ is a weight parameter used to balance the relative importance between distance and obstacle avoidance cost. A obstacle Represents obstacle areas in the environment, such as other plants, buildings, or any objects that the robot is not allowed to collide with; argmin P It means selecting the path that minimizes the objective function, that is, selecting a path with the shortest path and the lowest obstacle avoidance cost.

[0051] This formula ensures that the robot can complete the pruning task efficiently and accurately by minimizing the distance of the target path and avoiding obstacles as much as possible.

[0052] In this embodiment, step S3 uses the path planning module to generate an optimal path for the pruning robot based on the pruning plan. Path planning not only considers the efficiency of the pruning task but also ensures real-time obstacle avoidance to ensure continuity and accuracy during the pruning process. By comprehensively considering the robot's kinematics, environmental factors, and the priority of the pruning task, the system can flexibly adjust the robot's path in a dynamic environment, ensuring smooth pruning operations, thereby improving pruning efficiency and reducing manual intervention.

[0053] During the S4's execution of the pruning plan, the pruning robot is precisely controlled based on the optimal path. Specifically, during this step, the robot navigates along the generated optimal path while adjusting the force and angle of the pruning tool in real time to achieve the desired pruning effect.

[0054] In this embodiment, the pruning robot's path planning is based on multi-sensor data fusion technology. Sensors collect environmental information, plant status data, and obstacle locations, dynamically adjusting the robot's path to ensure efficient and accurate pruning. During this process, the robot's control system adjusts the force and angle of the pruning tool based on real-time feedback. This real-time feedback includes, but is not limited to, the force and angle of the pruning tool, the depth of pruning, and the contact between the tool and the plant.

[0055] Specifically, when implementing step S4, the pruning robot's control system needs to calculate the optimal pruning force and angle based on real-time feedback information, and automatically adjust the working state of the pruning tool according to the type, growth status, and environmental conditions of different plants. For example, when the sensor detects that the plant branches are relatively hard, the pruning tool's force may need to be increased, while when the plant leaves are detected to be relatively soft, the pruning tool's force should be appropriately reduced. Similarly, the pruning tool's angle also needs to be adjusted in real time based on the shape of the plant, the pruning location, and the target pruning effect.

[0056] During implementation, the control system of the pruning robot adjusts the force F and angle θ of the pruning tool in real time using the following mathematical formula: F = f(S, P, E); Among them, F represents the force of the pruning tool; S is the plant hardness data fed back by the sensor; P is the position and status of the plant; E is the environmental factors such as temperature, humidity and other factors affecting pruning.

[0057] θ=g(S,P,T); Where θ represents the angle of the pruning tool, S is the morphological data of the plant, P is the pruning target part, and T is the target type of the pruning task (such as pruning depth, angle requirements, etc.).

[0058] In the above formula, functions f and g can be specifically modeled and optimized according to different sensor data and environmental variables to ensure the smooth completion of the pruning task.

[0059] Typically, real-time feedback is transmitted to the central processing unit (CPU) via a communication protocol between the robot and the control system. The CPU then adjusts the robot's operations based on this real-time data. For example, if the robot detects that the trimming angle does not meet preset requirements, the control system can immediately instruct the robot to adjust the angle to ensure trimming accuracy.

[0060] As an option, the pruning robot's feedback information can also be used to monitor the working status of the pruning tools in real time through sensors, detect whether the pruning tools are worn or malfunctioning, and adjust the working status or issue an alarm in a timely manner to ensure the continuity and stability of the pruning operation.

[0061] Specifically, the control system can utilize multi-channel data fusion technology to integrate and analyze data from multiple sources, including visual sensors, tactile sensors, and accelerometers. Based on this comprehensive judgment, it can achieve optimal control of the force and angle of the pruning tool. For example, if the visual sensor detects that the pruning target position has shifted, the system can use computer vision algorithms to correct its position and adjust the force and angle based on the dynamic changes of the pruning target.

[0062] In one possible implementation, the pruning robot can also generate a personalized pruning plan based on real-time plant status feedback while executing the pruning plan. Within this plan, the robot automatically adjusts the force and angle of the pruning tool according to the growth status of each plant. For example, the robot may use different force and angle settings when pruning plants at different growth stages to achieve the best pruning effect and avoid excessive plant damage.

[0063] Furthermore, dynamic adjustments to the pruning path rely not only on static sensor data but also on real-time corrections based on the robot's movement pattern, speed, and changes in the external environment. These dynamic adjustments ensure that the pruning robot can maintain high pruning quality across different paths and environmental conditions.

[0064] In summary, this embodiment achieves efficient, accurate and automated pruning tasks through precise control of the force and angle of the pruning tool, and ensures continuous optimization and adjustment of the pruning process through real-time feedback information, avoiding pruning quality problems caused by unreasonable path planning and force setting.

[0065] The core goal is to dynamically adjust the pruning strategy based on the real-time data obtained during the pruning process to ensure that the final pruning results meet the expected standards, thereby achieving efficient, accurate and personalized pruning operations.

[0066] In step S5, the pruning strategy is optimized and adjusted by real-time monitoring of the pruning tool's operating status and the plant's response, combined with sensor feedback. During the pruning process, the pruning robot's control system continuously receives sensor data, such as plant hardness, branch angle, and the force and angle of the pruning tool, generating real-time feedback. This information is used to update the strategy for the current pruning task and automatically adjust parameters such as the pruning tool's force and angle based on the feedback to ensure continuous optimization of the pruning operation.

[0067] Generally speaking, the adjustment process of real-time feedback information can be carried out in the following ways: Adjusting the force of the pruning tool: When the robot receives feedback from sensors about the plant's hardness, the control system calculates the required pruning force in real time based on the plant's different parts (such as branches and leaves) and its growth status. If the plant is firm, the pruning force is increased; if the plant is soft, the force is reduced to minimize damage.

[0068] Adjusting the pruning angle: Based on sensor feedback about the plant's shape and the pruning target, the system can automatically adjust the angle of the pruning tool. For example, when pruning a relatively tilted branch, the control system will adjust the pruning angle to maintain consistent pruning depth and accuracy.

[0069] Specifically, when the control system makes adjustments based on real-time feedback information, the following mathematical formula can be used to achieve optimal adjustment of force and angle: F=h(H,A,T); Among them, F represents the force of the pruning tool; H is the plant hardness data fed back by the sensor; A is the plant morphological data (such as branch angle, position, etc.); T is the type of pruning task (such as pruning target depth, etc.).

[0070] θ=k(P,D,T); Where θ represents the angle of the pruning tool; P is the current position and state of the plant; D is the target pruning depth, and T is the task type (e.g., target pruning angle, etc.).

[0071] In the above formula, functions h and k represent the mapping relationship of real-time feedback information. They can be optimized according to the actual needs of the pruning target to ensure the efficiency and accuracy of the pruning process.

[0072] Alternatively, the system can further enhance the adaptability of pruning strategies by establishing a feedback control mechanism. For example, based on fuzzy control theory, the pruning system can adaptively adjust pruning force and angle based on highly uncertain plant data (such as changes in plant hardness or morphology) to cope with pruning tasks in complex environments.

[0073] In one possible implementation, real-time feedback can also help the pruning robot promptly detect and correct potential errors during pruning, such as excessive pruning depth or inaccurate angles. In these cases, the control system uses its internal self-calibration algorithm to make real-time corrections to the pruning task, recalculating pruning parameters and outputting a new pruning strategy. This not only improves pruning efficiency but also ensures consistent pruning quality.

[0074] Specifically, adjustments to pruning strategies rely not only on a single piece of feedback but also on the integration of multiple feedback sources. For example, the system combines data from multiple sensors, including visual sensors, pressure sensors, and displacement sensors, to construct a multi-dimensional pruning parameter adjustment model. This allows the system to automatically adjust parameters such as the force and angle of the pruning tool during the pruning process, avoiding inconsistent pruning caused by changes in the external environment or differences in plant growth.

[0075] As a possible extension technology, pruning robots could be combined with artificial intelligence algorithms to gradually learn and optimize pruning strategies. By analyzing historical pruning data, the robot can identify the optimal pruning strategy for different plant types and environmental conditions, thereby improving the automation and accuracy of future pruning tasks.

[0076] Finally, step S5 includes outputting the pruning results. By comparing the real-time data collected during the pruning process with the predetermined pruning targets, the system can evaluate the pruning effect and generate a final pruning result report. These reports include not only the state of the plant after pruning, but also the specific parameters used during the pruning process (such as force, angle, path, etc.), providing a basis for subsequent analysis and optimization.

[0077] In summary, this embodiment, through a dynamic adjustment mechanism based on real-time feedback information, can flexibly optimize pruning strategies during the pruning process, improving pruning quality and efficiency. At the same time, the system can promptly detect and resolve errors that may occur during the pruning process, ensuring the accuracy and stability of the pruning results.

[0078] The primary purpose of step S6 is to ensure that the pruning results meet expected standards through real-time evaluation. If deviations are detected during the evaluation process, the system automatically adjusts the pruning robot's optimal path or strategy to optimize the pruning effect. This step ensures the accuracy and effectiveness of the pruning operation, resulting in the best possible pruning results.

[0079] In this embodiment, step S6 performs a real-time assessment of the pruned plants, analyzing the pruning effectiveness based on feedback from multiple sensors. This assessment includes factors such as pruning depth, angle, and accuracy. The system compares the intended pruning target with the actual pruning results to identify potential deviations. For example, if the pruning angle is too large or the pruning depth is insufficient, the system will adjust the pruning robot's path or pruning strategy accordingly to optimize the pruning result.

[0080] Typically, a pruning robot's feedback system includes multiple sensors, such as vision sensors, force sensors, and angle sensors. These sensors provide information about the plant's condition after pruning, including surface quality, residual branches and leaves, and morphological changes at the cut site. By processing this information in real time, the system can identify potential deviations in the pruning process and adjust the robot's path or strategy accordingly.

[0081] For example, if the assessment finds that the trimming depth of a certain section is insufficient, the system can calculate the required trimming adjustment using the following formula: Δd=d target -d actual ; Among them, Δd is the trimming depth that needs to be adjusted; d target is the predetermined pruning depth; d actual is the actual trimming depth. Based on the calculated adjustment amount, the system will replan the robot's path or adjust the trimming strategy to ensure that the trimming meets the predetermined standards.

[0082] Specifically, the pruning robot can correct deviations by optimizing its path. For example, if the robot's path deviates, resulting in a missed pruning angle in certain areas, the control system adjusts the robot's optimal path based on post-pruning evaluation results. This adjustment is based on analysis of real-time feedback data, using a path planning algorithm to optimize the route, enabling the robot to accurately complete the pruning task.

[0083] In one possible implementation, the optimal path adjustment for a pruning robot can automatically generate a new path plan based on the difference between the pruning target and the actual pruning result. The control system combines the robot's current position, environmental data, pruning requirements, and evaluation feedback to calculate the new pruning path in real time. For example, if the robot deviates during the pruning process, the system can use a path algorithm to replan the robot's path based on the error adjustment formula, ensuring that the robot can prune the areas that did not meet the target, thereby compensating for the under-pruning.

[0084] Alternatively, this embodiment can employ machine learning to analyze deviations from historical pruning tasks and learn more precise pruning strategies. Using deep learning and other techniques, the pruning robot can continuously optimize its pruning paths and strategies during the execution of its tasks, achieving even greater accuracy.

[0085] Furthermore, real-time evaluation and deviation adjustment during the pruning process are not limited to pruning depth, but can also provide a comprehensive analysis of pruning angle, force, etc. For example, if the robot's pruning angle is too large or too small, resulting in uneven pruning, the system will use the formula to calculate the required angle adjustment based on real-time feedback information: Δθ=θ target -θ actual ; Among them, Δθ represents the trimming angle that needs to be adjusted; θ target is the target trimming angle, θ actual The actual trimming angle is obtained by the system. In this way, the system can accurately determine and correct the angle deviation.

[0086] In some embodiments, the real-time evaluation and deviation adjustment of the pruning robot after the pruning task is completed can be achieved through a feedback control mechanism. The feedback control mechanism uses the data collected by the sensor to compare with the target pruning data in real time, and then calculates the error in the pruning task. The deviation is corrected by adjusting the control parameters (such as pruning depth, force, angle, etc.) to ensure the pruning effect is consistent with the expectation.

[0087] In summary, this embodiment uses a real-time evaluation and correction mechanism after a pruning task is completed to automatically adjust the pruning robot's path or pruning strategy to optimize pruning results. This technical approach ensures the efficiency and accuracy of pruning operations through a feedback loop. With each pruning task, the system continuously accumulates data, enhancing the intelligence of the pruning process.

[0088] See also Figure 2 The present invention also provides a green plant intelligent pruning system, comprising: A sensor module for collecting real-time data of the plants to be pruned; Machine learning module for developing pruning plans based on real-time data; The path planning module is used to generate the optimal travel path of the pruning robot according to the pruning plan and adjust the path to meet the requirements of the pruning task; A control module, used for controlling the pruning robot to execute a pruning plan according to an optimal travel path; The feedback module is used to monitor the execution of the pruning plan of the pruning robot in real time, collect the force and angle of the pruning tool and the response data of the plant, and output the generated pruning results; The evaluation module is used to evaluate the pruning results after the pruning task is completed and adjust the pruning path or strategy to optimize the pruning effect.

[0089] Sensor module: This module is used to collect real-time data of the plants to be pruned. The data types include information such as the height, shape, density, distribution of branches and leaves, the state of the leaves, and the temperature and humidity of the surrounding environment. The sensor module usually includes visual sensors (such as cameras or laser scanners), force sensors, temperature and humidity sensors, etc. These sensors can capture the growth characteristics of plants and force feedback information during the pruning process with high precision. Depending on the application scenario, the type of sensor can be selectively optimized. For example, when pruning tall trees, a laser radar (LiDAR) sensor can be used to scan the crown structure, while when pruning shrubs or low plants, it may rely more on cameras and force sensors.

[0090] Machine learning module: This module uses machine learning algorithms (such as deep learning, reinforcement learning, etc.) based on real-time data collected by sensors to intelligently analyze the pruning needs of plants and formulate the optimal pruning plan. This module can continuously optimize pruning strategies by learning data patterns in historical pruning tasks. The machine learning module can not only process the real-time data of the current pruning task, but also adjust the pruning strategy based on the accumulated historical data, so that the pruning plan can be continuously optimized over multiple pruning cycles. Specifically, the machine learning module can conduct in-depth analysis of the plant's morphology, pruning targets, and real-time feedback data, and automatically calculate the optimal pruning angle, depth, and strength.

[0091] Path Planning Module: Based on the pruning plan developed by the Machine Learning Module, this module generates the optimal path for the pruning robot. This module considers the actual plant morphology, pruning objectives, and environmental obstacles to optimize the efficiency and safety of the robot's path. Furthermore, the path planning module is capable of real-time path adjustments to account for potential changes during the pruning task, such as changes in the plant life within the pruning area or deviations in the robot's position. The path planning module typically uses graph-based algorithms, A* algorithms, or dynamic programming algorithms to optimize the robot's path.

[0092] The control module receives the optimal path provided by the path planning module and precisely controls the pruning robot to execute the pruning plan. This module not only controls the robot's speed and steering but also coordinates the operation of the pruning tools, adjusting their force and angle. The control module responds to real-time feedback from the robot's sensors, ensuring that each pruning action is executed according to the plan. The control module typically includes motion control and motion optimization algorithms to ensure accurate robot movements.

[0093] Feedback Module: This module monitors the robot's execution of its pruning plan in real time. Sensors collect real-time data from the pruning tool, such as force, angle, and temperature, and assess the plant's response in real time, such as the degree of plant bending and the depth of pruning. Based on this real-time data, the feedback module generates feedback during the pruning process and provides it to the control and evaluation modules. The feedback module can also dynamically adjust pruning strategies during the pruning process to optimize accuracy and efficiency.

[0094] Evaluation Module: After a pruning task is completed, the evaluation module is responsible for real-time evaluation of pruning results to determine whether they meet the predetermined pruning objectives. If any deviation is detected, the evaluation module sends a signal to the path planning module or the control module to adjust the pruning robot's optimal path or pruning strategy to optimize pruning results. The evaluation module not only relies on visual feedback to assess pruning results, but also combines data from force and angle sensors to comprehensively assess pruning depth, angle, and their impact on plant health. This module continuously refines pruning plans based on these evaluation results, ensuring even more precise pruning next time.

[0095] See also Figure 3 The present invention also provides a green plant intelligent pruning robot, comprising: Vision sensors, lidar, near-infrared sensors, and force sensors to capture plant images, 3D morphology, health data, and the forces applied during pruning; The processing unit is used to formulate pruning plans based on real-time data and control the force, angle and path adjustment of the pruning tools.

[0096] Vision sensor: used to capture image data of plants to help analyze plant morphology, health status, and pruning needs. It supports plant feature extraction and growth status assessment, providing a basis for pruning plan formulation.

[0097] Laser radar (LiDAR): It acquires three-dimensional morphological data of plants by emitting laser beams, assisting robots in scanning the size and position of plants, avoiding obstacles, and providing depth information for path planning to generate clear plant outlines.

[0098] Near-infrared sensor: collects plant health data, detects physiological characteristics such as moisture and chlorophyll, identifies diseased or water-deficient areas, and ensures plant health is protected during pruning to avoid unnecessary damage.

[0099] Force sensor: This monitors the force applied by the pruning tool in real time, ensuring the appropriate force is applied during pruning to avoid damaging the plant. Force data is combined with other sensor information to adjust pruning depth, force, and angle.

[0100] The processing unit, the brain of the system, processes sensor data, develops pruning plans using machine learning and path planning algorithms, controls the force and angle of the pruning tool, and adjusts the path in real time to ensure optimal pruning results. The processing unit controls the force and angle of the pruning tool, ensuring real-time adjustments based on the plant's hardness and shape to avoid damage and achieve precise pruning. The robot also adjusts and optimizes the path by monitoring path deviations and obstacles in real time, ensuring smooth pruning. It also optimizes the path to improve efficiency and effectiveness, avoiding duplicate pruning.

[0101] The integration of these components enables the intelligent pruning robot of the present invention to complete the task of pruning green plants efficiently and accurately.

[0102] See also Figure 4 The present invention also provides a green plant intelligent pruning control device, comprising: Data acquisition module, used to collect real-time data of plants and transmit it to the machine learning module; The evaluation module is used to formulate a pruning plan based on the data and coordinate the path planning module to control the execution of the robot; Feedback module, used to provide real-time status of pruning tools and plant response, and output generated pruning results; The pruning control module is used to adjust the pruning task according to the pruning results.

[0103] Data Acquisition Module: This module is responsible for collecting real-time plant data, including images, 3D morphology, health status, and the force applied during pruning. This data is acquired through a variety of sensors, including visual sensors, lidar, near-infrared sensors, and force sensors. This data is transmitted in real time to the machine learning module, serving as the basis for subsequent pruning plans.

[0104] Evaluation Module: Based on collected data, the evaluation module uses machine learning algorithms to analyze the plant's growth status, health, and pruning needs, generating a pruning plan. Specifically, the evaluation module assesses the growth pattern and pruning requirements of each plant part and collaborates with the path planning module to control the robot's execution. The evaluation module also adjusts the pruning strategy based on actual conditions to ensure optimal pruning results.

[0105] Feedback Module: This module monitors the status of the pruning tool (e.g., force, angle, etc.) and the plant's response data (e.g., surface damage) in real time, and promptly outputs the pruning results. This module provides feedback on each step of the pruning process, enabling the control system to adjust the pruning strategy in real time. This feedback module ensures the accuracy and efficiency of the pruning process, avoiding damage to the plant or unsatisfactory pruning results.

[0106] Pruning Control Module: The Pruning Control Module is responsible for adjusting pruning tasks in real time based on the pruning results output by the Feedback Module. It can adjust the robot's path, the force, and the angle of the pruning tool based on the plant's response or changes in the external environment. Working in conjunction with other modules, the Pruning Control Module ensures dynamic adjustments to the pruning process after the task is completed, optimizing pruning results and minimizing disturbance to the plant during the pruning process.

[0107] The above modules work together to enable the intelligent pruning control device of the present invention to accurately perform plant pruning tasks with the support of real-time dynamic data, and optimize the pruning effect based on real-time feedback, ensuring the efficiency and accuracy of the plant pruning process.

[0108] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent pruning of green plants, characterized in that: The following steps are involved: Acquire real-time data of plants in the area to be pruned through a variety of sensors; Use machine learning models to evaluate real-time plant data to develop personalized pruning plans; generating an optimal travel path for the pruning robot based on the pruning plan; Controlling the pruning robot to execute the pruning plan according to the optimal travel path and providing real-time feedback information, wherein the real-time feedback information includes at least the force and angle of the pruning tool; During the pruning process, the pruning strategy is adjusted according to the real-time feedback information and the pruning results are output; After the pruning task is completed, the pruning results are evaluated in real time. If deviations are found, the optimal travel path or strategy of the pruning robot is adjusted to optimize the effect.

2. The intelligent pruning method for green plants according to claim 1, characterized in that: The various sensors include visual sensors, laser radars, and near-infrared sensors, which are used to obtain real-time data of plants in the area to be pruned.

3. The intelligent pruning method for green plants according to claim 1, characterized in that: The real-time data of the plants include morphological characteristics, health status and growth trends of the plants.

4. The intelligent pruning method for green plants according to claim 1, characterized in that: The machine learning model uses a deep learning algorithm to evaluate the morphological characteristics, health status and growth trends of plants by training historical pruning data, wherein the historical pruning data includes the morphological characteristics of plants, the health status of plants, pruning history records and pruning effects.

5. The intelligent pruning method for green plants according to claim 1, characterized in that: The pruning plan includes the following steps: Assess plant growth needs based on real-time plant data; Determine the pruning force, pruning angle and pruning position; Develop personalized pruning plans based on the plant's morphological characteristics, health status, and morphological features; Optimize and adjust each step of the pruning task.

6. The intelligent pruning method for green plants according to claim 1, characterized in that: The pruning results include: Morphological data of plants after pruning; Assessment of plant health status after pruning; Evaluation results of pruning accuracy and consistency; Deviations that occur during pruning and corresponding pruning adjustment measures.

7. The intelligent pruning method for green plants according to claim 1, characterized in that: The real-time assessment includes: Obtain plant morphology and health status data after pruning; Compare plant data before and after pruning to evaluate pruning accuracy and effectiveness; Detect deviations in the pruning process and analyze them; Adjust the pruning path or strategy based on the evaluation results to ensure optimal pruning effect.

8. A green plant intelligent pruning system, applied to the green plant intelligent pruning method according to any one of claims 1 to 7, characterized in that: include: A sensor module for collecting real-time data of the plants to be pruned; Machine learning module for developing pruning plans based on real-time data; The path planning module is used to generate the optimal travel path of the pruning robot according to the pruning plan and adjust the path to meet the requirements of the pruning task; A control module, used for controlling the pruning robot to execute a pruning plan according to an optimal travel path; The feedback module is used to monitor the execution of the pruning plan of the pruning robot in real time, collect the force and angle of the pruning tool and the response data of the plant, and output the generated pruning results; The evaluation module is used to evaluate the pruning results after the pruning task is completed and adjust the pruning path or strategy to optimize the pruning effect.

9. A green plant intelligent pruning robot, applied to the green plant intelligent pruning method according to any one of claims 1 to 7, characterized in that: include: Vision sensors, lidar, near-infrared sensors, and force sensors to capture plant images, 3D morphology, health data, and the forces applied during pruning; The processing unit is used to formulate pruning plans based on real-time data and control the force, angle and path adjustment of the pruning tools.

10. A green plant intelligent pruning control device, applied to the green plant intelligent pruning method according to any one of claims 1 to 7, characterized in that: include: Data acquisition module, used to collect real-time data of plants and transmit it to the machine learning module; The evaluation module is used to formulate a pruning plan based on the data and coordinate the path planning module to control the execution of the robot; Feedback module, used to provide real-time status of pruning tools and plant response, and output generated pruning results; The pruning control module is used to adjust the pruning task according to the pruning results.

Images