Lawn autonomous maintenance management method, device, equipment, medium and system

Abstract

The invention relates to the technical field of smart gardens, and discloses a lawn autonomous maintenance management method, device, equipment, medium and system.The lawn health semantic map representing the lawn health condition is built and updated in real time, and when it is monitored that the health state of any grid unit in the map is changed from normal to abnormal, the health state of any grid unit in the map is monitored to be abnormal; and generating a maintenance task containing the position and the problem type, and scheduling the maintenance task to corresponding execution equipment according to a preset rule. According to the method, lawn health state recognition is carried out by fusing multi-mode sensing data and utilizing a multi-branch neural network model which is subjected to progressive and adversarial training. Meanwhile, a state verification closed loop after task execution is introduced, and the maintenance effect is evaluated. According to the invention, full-process closed-loop automatic management from state sensing, intelligent decision making, accurate execution to effect verification can be realized, the problems of link splitting, response lag, resource waste and experience dependence in the prior art are solved, and the lawn maintenance accuracy, the intelligent level and the resource utilization efficiency are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of smart garden technology, and in particular to a method, device, equipment, medium and system for autonomous lawn maintenance and management. Background Technology

[0002] Modern landscaping, especially lawns in large public green spaces, sports fields, golf courses, and courtyards, involves a variety of complex operations such as watering, fertilizing, spraying pesticides, weeding, and replanting. Currently, the industry's commonly used lawn maintenance management model still heavily relies on manual experience for planning, scheduling, and decision-making. Each stage is relatively independent, resulting in the following major fundamental flaws: 1. The lawn health status monitoring process is completely separated from the actual maintenance operations in terms of workflow and system. The monitoring results are usually output in the form of offline reports or static images, requiring experienced managers to manually interpret and judge them, and then manually formulate maintenance plans and dispatch personnel and equipment. This serialized process with multiple human interventions results in a lengthy cycle from problem discovery to the formation of an executable task, which is prone to misunderstandings and delays in information transmission, and overall decision-making and response efficiency is low.

[0003] 2. Traditional lawn maintenance often employs a uniform operation model based on fixed cycles or manual experience, such as uniform irrigation, fertilization, and pesticide application throughout the lawn. This approach ignores the spatial and temporal differences in the lawn's internal microenvironment and health status, leading to significant waste of water, fertilizer, and pesticide resources in healthy areas that require no treatment, while potentially insufficient supplies are provided to problem areas that truly need intensive maintenance. This model not only significantly increases maintenance costs but also causes resource waste and potential environmental pollution, contradicting the development concepts of precision agriculture and green maintenance.

[0004] 3. Existing solutions generally lack effective tracking and recording of the maintenance task execution process, and also lack objective and quantitative verification and evaluation methods for the effects of maintenance measures. Managers often cannot monitor the location, status, and operating parameters of equipment in real time, and it is even more difficult to correlate maintenance operations in specific areas with subsequent changes in their health status. This leaves maintenance management largely reliant on personal experience, making it difficult to form a data-driven closed loop of "execution-evaluation-optimization," and hindering the scientific evaluation of the effectiveness of maintenance strategies and continuous improvement.

[0005] In summary, existing lawn maintenance solutions suffer from systemic flaws such as a disconnect between detection and execution, passive and delayed response, extensive and wasteful operations, and reliance on experience for management. The various stages are fragmented and fail to form a collaborative and intelligent organic whole. Therefore, there is an urgent need for a solution that can connect the entire process from status perception, intelligent decision-making, precise execution to effect verification, in order to achieve a precise, automated, and intelligent upgrade of autonomous lawn maintenance management. Summary of the Invention

[0006] This invention provides a method, device, equipment, medium, and system for autonomous lawn maintenance and management, in order to overcome the shortcomings of the prior art.

[0007] This invention provides a method for autonomous lawn maintenance and management, comprising: The update status of the lawn health semantic map of the target lawn area is monitored. The lawn health semantic map divides the target lawn area into multiple grid cells, and each grid cell is identified in real time with lawn health status information and area location information. When the lawn health status information of a certain grid cell on the lawn health semantic map changes from normal to abnormal, a maintenance task is generated based on the new lawn health status information of the grid cell and the regional location information. According to the preset task-device mapping rules, the maintenance task is sent to the corresponding task execution device.

[0008] According to the present invention, a lawn self-maintenance and management method is provided, wherein the lawn health semantic map is a multi-layer raster data stack, and each raster unit occupies the same spatial position in all layers of the multi-layer raster data stack. The multi-layer raster data stack includes a geographic layer, a semantic layer, and an extended information layer. The geographic layer includes a coordinate layer and an elevation layer. The semantic layer includes a lawn health status label layer, a confidence layer, and a timestamp layer. The extended information layer includes a vegetation index layer and a texture layer.

[0009] The lawn self-maintenance and management method provided by the present invention further includes: Obtain the lawn health status identification result in the target lawn area, wherein the lawn health status identification result includes the current lawn health status, the confidence level of the current identification result, the current area location information, and the current timestamp; Obtain the set of grid cells corresponding to the current area location information from the lawn health semantic map; Based on the lawn health status recognition results, the layer information of each grid cell in the grid cell set is updated according to the preset data update rules to obtain the updated lawn health semantic map.

[0010] According to the present invention, a method for autonomous lawn maintenance and management, wherein obtaining the lawn health status identification result in the target lawn area includes: Obtain raw sensing data of the target lawn area, wherein the raw sensing data includes at least one of RGB image blocks, elevation blocks, vegetation index blocks, and real-time vibration data; Based on the original sensing data, the lawn health status recognition result of the target lawn area is obtained through a preset lawn health status recognition model.

[0011] According to a method for autonomous lawn maintenance and management provided by the present invention, before acquiring the original sensing data of the area to be tested, the method further includes: Construct the preset lawn health status recognition model; Specifically, historical RGB image patches, historical elevation patches, historical vegetation index patches, historical vibration data, and health status labels of the lawn area are obtained. The health status labels include coarse-grained health status labels and problem type labels. Based on the historical RGB image patches, historical elevation patches, historical vegetation index patches, historical vibration data, and the health status labels, a multi-branch neural network architecture is used, employing a progressive training strategy and an adversarial mode dropout training method to train the model to learn the feature changes of the historical RGB image patches, historical elevation patches, historical vegetation index patches, and historical vibration data under different health states, thereby obtaining the lawn health status recognition model.

[0012] According to the present invention, a method for autonomous lawn maintenance and management is provided, wherein the multi-branch neural network architecture includes: The visual branch is used to extract visual feature maps of different levels from the historical RGB image blocks and perform visual feature fusion to obtain multi-scale visual feature vectors. The geometric branch is used to extract geometric feature maps of different levels from the historical elevation map tiles and perform geometric feature fusion to obtain multi-scale geometric feature vectors. The spectral branch is used to extract spectral feature maps of different levels from the historical vegetation index map and perform spectral feature fusion to obtain a multi-scale spectral feature vector. The physical branch is used to extract multiple vibration features from the historical vibration data and fuse the vibration features to obtain a physical feature vector. The adaptive branch weight learning module is used to predict the confidence score of each branch on the current sample, calculate the adaptive weight of each branch, and process the output vector of each branch using the adaptive weight of each branch to obtain the weighted vector of each branch. Cross-modal attention fusion network is used to enhance and fuse the weighted vectors of each branch using a cross-modal attention mechanism to obtain a multimodal feature fusion vector; A multi-granularity classifier is used to output the lawn health status recognition result based on the fusion vector of multimodal features.

[0013] According to the present invention, a method for autonomous lawn maintenance and management includes a progressive training strategy comprising: Single-modal pre-training phase: Pre-training of the corresponding branch is performed using training data from only one of the four modalities: visual, geometric, spectral, and physical. Bimodal fusion training phase: Bimodal fusion training is performed using training data from two of the four modalities: visual, geometric, spectral, and physical. Full-modal fusion training phase: Combining training data from four modalities—visual, geometric, spectral, and physical—adaptive branch weight learning module and cross-modal attention fusion network are enabled to conduct full-modal fusion training; Multi-granularity classification training stage: Combining training data from four modalities—visual, geometric, spectral, and physical—and utilizing a multi-task loss function, a main classifier, a fine-grained classifier, and a confidence output layer are trained. The main classifier outputs whether the lawn area to be treated is healthy, the fine-grained classifier outputs the types of problems existing in the lawn area to be treated, and the confidence output layer outputs the model's confidence in the prediction results.

[0014] According to the present invention, the adversarial modal discarding training method refers to randomly discarding any number of training data of any number of modalities during the training process of the lawn health status recognition model, so as to simulate the situation of data missing and / or lost.

[0015] According to the present invention, a method for autonomous lawn maintenance and management includes the following preset data update rules: If the original lawn health status is healthy or unknown, or if the difference between the confidence level of the current lawn health status identification result and the confidence level of the original lawn health status identification result is greater than or equal to the first preset value, the original lawn health status identification result is updated with the current lawn health status identification result. If the current lawn health status is the same as the original lawn health status, update the timestamp; If the current lawn health status is different from the original lawn health status, and the difference between the confidence level of the current lawn health status identification result and the confidence level of the original lawn health status identification result is less than a second preset value, a conflict negotiation mechanism is triggered. The conflict negotiation mechanism includes any one of the following or any combination thereof: retaining the lawn health status label with higher confidence, marking the lawn health status label as pending confirmation, or making a decision based on the logical priority of the problem type.

[0016] The lawn self-maintenance and management method provided by the present invention further includes: Obtain the status and location information of the task execution device; Based on the status information and the location information, combined with the maintenance task and the preset task priority sorting rules, an optimal task sequence and an optimal movement path are generated for the task execution device.

[0017] According to the present invention, the status information of the task execution device includes power and / or material loading.

[0018] The lawn self-maintenance and management method provided by the present invention further includes: Receive task completion information from the task execution device and mark the maintenance task as pending verification on the lawn health semantic map; The update status of the lawn health semantic map is continuously monitored. When the lawn health status information of the area corresponding to the maintenance task changes from abnormal to normal, the maintenance task is determined to be executed successfully. When the lawn health status information of the area corresponding to the maintenance task remains abnormal, the maintenance task is determined to be executed unsuccessfully. If the maintenance task is determined to have failed, the maintenance task will be regenerated or an alarm will be triggered.

[0019] The present invention also provides a lawn self-maintenance and management device, comprising: The monitoring module is used to monitor the update status of the lawn health semantic map of the target lawn area. The lawn health semantic map divides the target lawn area into multiple grid cells, and each grid cell is identified in real time with lawn health status information and area location information. The task generation module is used to generate a maintenance task based on the new lawn health status information and regional location information of the grid cell when the lawn health status information of a certain grid cell on the lawn health semantic map changes from normal to abnormal. The task sending module is used to send maintenance tasks to the corresponding task execution devices according to preset task-device mapping rules.

[0020] The present invention also provides an electronic device, including a processor and a memory storing a computer program, wherein the processor executes the computer program to implement any of the above-described lawn autonomous maintenance and management methods.

[0021] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements any of the above-described methods for autonomous lawn maintenance and management.

[0022] The present invention also provides a computer program product, the computer program product comprising a computer program that can be stored on a non-transitory computer-readable storage medium, and when the computer program is executed by a processor, the computer is able to execute any of the above-described lawn autonomous maintenance and management methods.

[0023] The present invention also provides a lawn autonomous maintenance management system for implementing any of the above-described lawn autonomous maintenance management methods, comprising an inspection device, a server, and a task execution device. The inspection device is used to inspect the target lawn area and update the lawn health status information and area location information of each grid cell on the lawn health semantic map in real time. The server is used to monitor the update status of the lawn health semantic map, generate maintenance tasks, and send the maintenance tasks to the corresponding task execution device. The task execution device is used to receive the maintenance tasks sent by the server and execute them.

[0024] The present invention provides a method, apparatus, equipment, medium, and system for autonomous lawn maintenance and management, which has at least the following beneficial effects: This invention fundamentally changes the traditional sequential process of separating detection and execution and relying on manual decision-making by constructing and maintaining a dynamically updated semantic map of lawn health in real time, and establishing an automatic triggering mechanism for its state changes and maintenance task generation. When the health status of any grid cell in the map changes from normal to abnormal, a maintenance task containing precise location and problem type can be automatically generated immediately and automatically scheduled to the corresponding equipment according to preset rules. This enables seamless connection and automation of the entire process from "perception-analysis-decision-execution", shortening the long cycle from problem discovery to task formation in the traditional model to near real-time, significantly improving the response speed to lawn problems and the level of intelligence in decision-making and execution.

[0025] The lawn health semantic map of this invention subdivides the target area into numerous grid units, each carrying independent health status information. Maintenance tasks are generated and executed based on "problem grids," enabling differentiated and precise operations that address needs only where they are required. This allows for precise control of the amount and location of water, fertilizer, and pesticides applied, avoiding waste in healthy areas and insufficient application in problem areas, significantly reducing maintenance material costs, and minimizing the negative environmental impact of overuse—perfectly aligning with the core concepts of precision agriculture and green maintenance.

[0026] This invention constructs and continuously updates a semantic map of lawn health using multimodal sensing data, comprehensively recording the historical evolution of lawn status. It records maintenance decisions and execution processes through automatic task generation and equipment scheduling. More importantly, by introducing a post-hoc closed loop of "task execution - status verification," it can automatically monitor whether the health status of the corresponding area has recovered after the maintenance task is executed. If the abnormality persists, the task is deemed a failure and an alarm or reassignment can be triggered. This mechanism allows the effectiveness of each maintenance action to be objectively and quantitatively verified, thus forming a data-driven closed-loop management system of "status monitoring → task execution → effect evaluation → strategy optimization." This makes maintenance management no longer dependent on personal experience, but based on traceable and analyzable data, enabling scientific evaluation and continuous optimization of maintenance strategies.

[0027] The lawn health status recognition model employed in this invention, based on a multi-branch neural network architecture, integrates four modalities: visual (RGB), geometric (elevation), spectral (vegetation index), and physical (vibration). Through progressive training and adversarial modality dropout training strategies, the model possesses powerful feature learning and generalization capabilities, as well as stable inference capabilities even with partial data loss. Adaptive branch weight learning and cross-modal attention fusion mechanisms dynamically adjust the contribution of each modality's information according to specific scenarios, achieving more accurate status judgments. The multi-granularity classifier not only determines "whether it is healthy" but also identifies specific "problem types" and outputs confidence scores, providing highly reliable data input for subsequent task generation and conflict negotiation.

[0028] This invention not only automatically generates tasks but also further considers equipment status (power level, material load), location, and task priority to generate optimal task sequences and paths for the equipment, thereby optimizing resource utilization and operational efficiency. Simultaneously, it introduces a conflict negotiation mechanism based on confidence comparison and logical priority into the lawn health semantic map update rules. This mechanism can intelligently handle inconsistencies between old and new detection results, avoiding false triggering of tasks due to single false detections or data noise, improving the robustness and fault tolerance of decision-making, and ensuring the long-term, stable, and reliable operation of the entire automated system. Attached Figure Description

[0029] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0030] Figure 1 This is a flowchart illustrating a method for autonomous lawn maintenance and management provided by the present invention.

[0031] Figure 2 This is a schematic diagram of the structure of a lawn self-maintenance and management device provided by the present invention.

[0032] Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, embodiments of this invention, and should not be construed as limiting the invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention. In the description of this invention, it should be understood that the terminology used is for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0034] Figure 1 This is a flowchart illustrating a method for autonomous lawn maintenance and management provided by the present invention. The executing entity of this method can be any applicable terminal-side device or network-side device, such as a yard management terminal / server, a smart lawnmower robot, etc.

[0035] See Figure 1 The present invention provides a method for autonomous lawn maintenance and management, which may include: S110. Construct a preset lawn health status recognition model.

[0036] In one embodiment, a preset lawn health status recognition model can be constructed through the following steps: The historical RGB image patches, historical elevation patches, historical vegetation index patches, historical vibration data, and health status labels of the lawn area are obtained. The health status labels include coarse-grained health status labels and problem type labels. Based on the historical RGB image patches, historical elevation patches, historical vegetation index patches, historical vibration data, and the health status labels, a multi-branch neural network architecture is used, employing a progressive training strategy and an adversarial mode dropout training method to train the model to learn the feature changes of the historical RGB image patches, historical elevation patches, historical vegetation index patches, and historical vibration data under different health states, thereby obtaining the lawn health status recognition model.

[0037] The multi-branch neural network architecture includes: The visual branch is used to extract visual feature maps of different levels from the historical RGB image blocks and perform visual feature fusion to obtain multi-scale visual feature vectors. The geometric branch is used to extract geometric feature maps of different levels from the historical elevation map tiles and perform geometric feature fusion to obtain multi-scale geometric feature vectors. The spectral branch is used to extract spectral feature maps of different levels from the historical vegetation index map and perform spectral feature fusion to obtain a multi-scale spectral feature vector. The physical branch is used to extract multiple vibration features from the historical vibration data and fuse the vibration features to obtain a physical feature vector. The adaptive branch weight learning module is used to predict the confidence score of each branch on the current sample, calculate the adaptive weight of each branch, and process the output vector of each branch using the adaptive weight of each branch to obtain the weighted vector of each branch. Cross-modal attention fusion network is used to enhance and fuse the weighted vectors of each branch using a cross-modal attention mechanism to obtain a multimodal feature fusion vector; A multi-granularity classifier is used to output the lawn health status recognition result based on the fusion vector of multimodal features.

[0038] The progressive training strategy includes: Single-modal pre-training phase: Pre-training of the corresponding branch is performed using training data from only one of the four modalities: visual, geometric, spectral, and physical. Bimodal fusion training phase: Bimodal fusion training is performed using training data from two of the four modalities: visual, geometric, spectral, and physical. Full-modal fusion training phase: Combining training data from four modalities—visual, geometric, spectral, and physical—adaptive branch weight learning module and cross-modal attention fusion network are enabled to conduct full-modal fusion training; Multi-granularity classification training stage: Combining training data from four modalities—visual, geometric, spectral, and physical—and utilizing a multi-task loss function, a main classifier, a fine-grained classifier, and a confidence output layer are trained. The main classifier outputs whether the lawn area to be treated is healthy, the fine-grained classifier outputs the types of problems existing in the lawn area to be treated, and the confidence output layer outputs the model's confidence in the prediction results.

[0039] The adversarial modality dropout training method refers to randomly dropping any number of training data for any number of modalities during the training process of the lawn health status recognition model, in order to simulate the situation of data missing and / or lost.

[0040] S120. Obtain the lawn health status identification result in the target lawn area, wherein the lawn health status identification result includes the current lawn health status, the current identification result confidence level, the current area location information, and the current timestamp.

[0041] In one embodiment, S120 may include: Obtain raw sensing data of the target lawn area, wherein the raw sensing data includes at least one of RGB image blocks, elevation blocks, vegetation index blocks, and real-time vibration data; Based on the original sensing data, the lawn health status recognition result of the target lawn area is obtained through a preset lawn health status recognition model.

[0042] S130. Obtain a set of raster cells corresponding to the current area location information from the lawn health semantic map. Based on the lawn health status recognition result, update the layer information of each raster cell in the raster cell set according to a preset data update rule to obtain an updated lawn health semantic map. The lawn health semantic map is a multi-layer raster data stack, where each raster cell occupies the same spatial position in all layers of the multi-layer raster data stack. The multi-layer raster data stack includes a geographic layer, a semantic layer, and an extended information layer. The geographic layer includes a coordinate layer (used to store the latitude and longitude or planar coordinates of the center point of each raster cell) and an elevation layer (used to store the ground elevation value of each raster cell). The semantic layer includes a lawn health status label layer (the initial label value is "unknown" or "healthy," and is updated to a specific problem, such as "groundwater accumulation," when a recognition result is received). The extended information layer includes a vegetation index layer (used to store the vegetation index of each grid cell) and a texture layer (used to store the texture feature value of each grid cell). The extended information layer includes a confidence layer (initial value is 0), a timestamp layer (used to record the time when the label of each grid cell was last updated or confirmed), and a vegetation index layer (used to store the vegetation index of each grid cell) and a texture layer (used to store the texture feature value of each grid cell).

[0043] In one embodiment, the preset data update rule includes: If the original lawn health status is healthy or unknown, or if the difference between the confidence level of the current lawn health status identification result and the confidence level of the original lawn health status identification result is greater than or equal to the first preset value, the original lawn health status identification result is updated with the current lawn health status identification result. If the current lawn health status is the same as the original lawn health status, update the timestamp; If the current lawn health status is different from the original lawn health status, and the difference between the confidence level of the current lawn health status identification result and the confidence level of the original lawn health status identification result is less than a second preset value, a conflict negotiation mechanism is triggered. The conflict negotiation mechanism includes any one of the following or any combination thereof: retaining the lawn health status label with higher confidence, marking the lawn health status label as pending confirmation, or making a decision based on the logical priority of the problem type.

[0044] S140. Monitor the update status of the lawn health semantic map of the target lawn area, wherein the lawn health semantic map divides the target lawn area into multiple grid cells, and each grid cell is identified in real time with lawn health status information and area location information; when the lawn health status information of a certain grid cell on the lawn health semantic map changes from normal (healthy) to abnormal (unhealthy, with problems such as groundwater accumulation, brown spot disease), a maintenance task is generated based on the new lawn health status information and area location information of the grid cell; according to the preset task-device mapping rules, the maintenance task is sent to the corresponding task execution device.

[0045] In one embodiment, a preset task-device mapping rule can associate a problem type with a task execution device, for example: "Drought" -> "Automatic Irrigation System" "Nutrition deficiency" -> "Precision fertilization machine" "Groundwater" -> "Drilling machine" "Disease" -> "Precision pesticide application drone" S150. Obtain the status information (including power and / or load) and location information of the task execution device, and combine the maintenance task with the preset task priority sorting rules (e.g., the priority of maintenance tasks that deal with the spread of disease is higher than the priority of maintenance tasks that deal with local bald spots) to generate an optimal task sequence and an optimal movement path for the task execution device.

[0046] Specifically, path planning algorithms (such as a variant of the Traveling Salesman Problem (TSP)) can be used to generate an optimal movement path for the task execution device with the goal of minimizing the total operation time or the total travel distance.

[0047] Furthermore, when the area corresponding to a maintenance task is larger than a preset range, the maintenance task can be divided into multiple sub-tasks, which are then completed jointly by multiple task execution devices. The preset range can be pre-set based on the real-time working capacity of the task execution devices or dynamically set in real time. The real-time working capacity can be predicted based on the power consumption of the task execution devices.

[0048] S160. Receive task completion information (including the executed task ID, location, completion time, and possible material usage) from the task execution device. Mark the maintenance task as pending verification on the lawn health semantic map. Continuously monitor the update status of the lawn health semantic map. When the lawn health status information of the area corresponding to the maintenance task changes from abnormal to normal, determine that the maintenance task has been successfully executed (at this time, the task can be closed, and a successful "measure-result" pair can be recorded for subsequent model optimization). When the lawn health status information of the area corresponding to the maintenance task remains abnormal, determine that the maintenance task has failed. When the maintenance task is determined to have failed, regenerate the maintenance task or trigger an alarm. The working efficiency of the task execution device can also be evaluated based on the verification results.

[0049] The present invention provides a method for autonomous lawn maintenance and management, which has the following significant beneficial effects: By connecting the entire process from "perception-decision-execution-verification", it can automatically complete problem identification, task generation, equipment scheduling and effect evaluation, significantly reducing the need for manual intervention, improving operational efficiency and saving labor costs.

[0050] This transforms the traditional extensive "area-like maintenance" into "point-like maintenance" based on refined grid units, enabling precise on-demand delivery of resources such as water, fertilizer, and pesticides. While ensuring maintenance effectiveness, it significantly saves materials, reduces operating costs, and minimizes environmental pollution.

[0051] It fully records the entire chain of data on lawn condition evolution, maintenance task execution, and effect feedback, constructing a traceable, quantifiable, and analyzable data closed loop. This upgrades management decisions from relying on personal experience to scientific decisions based on objective data, supporting continuous optimization of maintenance strategies.

[0052] Based on a real-time updated semantic map of lawn health and an automatic task triggering mechanism, maintenance tasks can be generated and dispatched immediately after an abnormal state is detected, enabling rapid response and precise intervention to sudden lawn problems, effectively preventing the spread of problems and reducing repair costs.

[0053] Through standardized task interfaces and scheduling mechanisms, it can be compatible with multiple types and brands of intelligent maintenance equipment, forming a flexible and scalable collaborative network of maintenance equipment, providing system support for the large-scale and intelligent upgrading of lawn maintenance.

[0054] The lawn autonomous maintenance and management device provided by the present invention is described below. The lawn autonomous maintenance and management device described below can be referred to in correspondence with the lawn autonomous maintenance and management method described above.

[0055] See Figure 2The present invention provides a lawn self-maintenance and management device, which may include: The monitoring module is used to monitor the update status of the lawn health semantic map of the target lawn area. The lawn health semantic map divides the target lawn area into multiple grid cells, and each grid cell is identified in real time with lawn health status information and area location information. The task generation module is used to generate a maintenance task based on the new lawn health status information and regional location information of the grid cell when the lawn health status information of a certain grid cell on the lawn health semantic map changes from normal to abnormal. The task sending module is used to send maintenance tasks to the corresponding task execution devices according to preset task-device mapping rules.

[0056] In one embodiment, it further includes: The identification module is used to obtain the identification result of the lawn health status in the target lawn area, wherein the lawn health status identification result includes the current lawn health status, the confidence level of the current identification result, the current area location information, and the current timestamp; The extraction module is used to obtain a set of raster cells corresponding to the current area location information from the lawn health semantic map; The update module is used to update the layer information of each grid cell in the grid cell set according to the lawn health status identification result and a preset data update rule, so as to obtain the updated lawn health semantic map.

[0057] In one embodiment, the identification module includes: The test data acquisition module is used to acquire the raw sensing data of the target lawn area, wherein the raw sensing data includes at least one of RGB image blocks, elevation blocks, vegetation index blocks, and real-time vibration data; The status recognition module is used to obtain the lawn health status recognition result of the test area based on the original sensing data and through a preset lawn health status recognition model.

[0058] In one embodiment, it further includes: The model building module is used to build the preset lawn health status recognition model; Specifically, historical RGB image patches, historical elevation patches, historical vegetation index patches, historical vibration data, and health status labels of the lawn area are obtained. The health status labels include coarse-grained health status labels and problem type labels. Based on the historical RGB image patches, historical elevation patches, historical vegetation index patches, historical vibration data, and the health status labels, a multi-branch neural network architecture is used, employing a progressive training strategy and an adversarial mode dropout training method to train the model to learn the feature changes of the historical RGB image patches, historical elevation patches, historical vegetation index patches, and historical vibration data under different health states, thereby obtaining the lawn health status recognition model.

[0059] In one embodiment, the preset data update rule includes: If the original lawn health status is healthy or unknown, or if the difference between the confidence level of the current lawn health status identification result and the confidence level of the original lawn health status identification result is greater than or equal to the first preset value, the original lawn health status identification result is updated with the current lawn health status identification result. If the current lawn health status is the same as the original lawn health status, update the timestamp; If the current lawn health status is different from the original lawn health status, and the difference between the confidence level of the current lawn health status identification result and the confidence level of the original lawn health status identification result is less than a second preset value, a conflict negotiation mechanism is triggered. The conflict negotiation mechanism includes any one of the following or any combination thereof: retaining the lawn health status label with higher confidence, marking the lawn health status label as pending confirmation, or making a decision based on the logical priority of the problem type.

[0060] In one embodiment, it further includes: The device data acquisition module is used to acquire the status information and location information of the task execution device; The planning module is used to generate a preferred task sequence and a preferred movement path for the task execution equipment based on the status information and the location information, combined with the maintenance task and the preset task priority sorting rules.

[0061] In one embodiment, it further includes: The feedback information receiving module is used to receive the task completion information fed back by the task execution device and mark the maintenance task as pending verification on the lawn health semantic map; The task verification module is used to continuously monitor the update status of the lawn health semantic map. When the lawn health status information of the area corresponding to the maintenance task changes from abnormal to normal, the maintenance task is determined to be executed successfully. When the lawn health status information of the area corresponding to the maintenance task remains abnormal, the maintenance task is determined to be executed unsuccessfully. The alarm module is used to regenerate the maintenance task or trigger an alarm when the maintenance task is determined to have failed.

[0062] The present invention also provides a lawn autonomous maintenance management system for implementing the lawn autonomous maintenance management method described in any of the above claims, comprising an inspection device, a server, and a task execution device (which may be an autonomous mobile device, such as a lawn mowing robot). The inspection device is used to inspect the target lawn area and update the lawn health status information and area location information of each grid cell on the lawn health semantic map in real time. The server is used to monitor the update status of the lawn health semantic map, generate maintenance tasks, and send the maintenance tasks to the corresponding task execution device. The task execution device is used to receive the maintenance tasks sent by the server and execute them.

[0063] Example 1: Fully Automated Disease Prevention and Control for Golf Courses Process: In the early morning, a routine inspection drone detected early symptoms of brown spot disease in Zone 5 of the fairway and updated the lawn health semantic map in real time. The server detected the changes in the lawn health semantic map, immediately generated a high-priority "application task," and dispatched a nearby precision application drone. The precision application drone received the maintenance task and precise boundary coordinates, flew to the target area, completed targeted low-altitude spraying within 5 minutes, and reported task completion to the server. That afternoon, the inspection drone conducted another inspection, verifying that the disease was under control, and updated the lawn health semantic map. The server recorded the task as successfully completed. The entire process required no human intervention.

[0064] Example 2: Collaborative Water and Fertilizer Management in Parks Process: The lawn health semantic map shows area A as "drought" and area B as "nutrient deficient". The server generates two tasks: irrigation and fertilization. The scheduling engine finds that both automatic irrigation trucks and precision fertilizer applicators are available, but their operation paths overlap. Dynamic planning is activated to allow the irrigation truck to complete the operation in area A first, followed by the fertilizer applicator entering area B, avoiding equipment conflicts and path duplication. The entire park's water and fertilizer management is completed automatically at night.

[0065] Example 3: Integrated Weed Management in Large Estates Process: The intelligent lawnmower identifies multiple "weed" areas and updates the lawn health semantic map in real time. The server detects changes in the lawn health semantic map and generates "weeding tasks." Because the weeds are scattered, the server uses a scheduling engine to package multiple scattered tasks, plan an optimal path, and assign it to a weeding robot. The weeding robot travels along the planned path, activating its robotic arm or precisely spraying micro-dosage only at locations with weeds, greatly saving energy and herbicide.

[0066] Example 4: Maintenance Performance Assessment and Supplier Management Process: Server records show that the "underground water accumulation" task assigned to "Drill Machine A" achieved a success rate of 95% in subsequent verifications; while the success rate of the task assigned to "Drill Machine B" was only 70%. Based on this, the manager can determine that Drill Machine B may have mechanical failures or operational problems, and thus carry out targeted maintenance, achieving data-driven equipment performance management.

[0067] Example 5: Emergency Response and Resource Allocation Process: After the thunderstorm, the lawn health semantic map showed "water accumulation" in several low-lying areas. The server generated multiple "drainage warning" tasks. Although the priority was not high, the scheduling engine detected that a maintenance vehicle with pumping capabilities was performing other tasks in the park. It was then sent a detour instruction to the vehicle, allowing it to handle the two most severely waterlogged areas on its return journey, demonstrating the flexibility of dynamic scheduling.

[0068] Figure 3 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 3 As shown, the electronic device may include a processor 810, a communications interface 820, a memory 830, and a communication bus 840, wherein the processor 810, the communications interface 820, and the memory 830 communicate with each other via the communication bus 840. The processor 810 can call logical instructions in the memory 830 to execute the steps of the lawn autonomous maintenance and management method described above.

[0069] Furthermore, the logical instructions in the aforementioned memory 830 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) 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 USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0070] On the other hand, the present invention also provides a computer program product, the computer program product including a computer program, the computer program being stored on a non-transitory computer-readable storage medium, and when the computer program is executed by a processor, the computer is able to perform the steps of the lawn autonomous maintenance and management method described in any of the above claims.

[0071] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the lawn autonomous maintenance and management method described in any of the preceding claims.

[0072] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0073] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0074] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for autonomous lawn maintenance and management, characterized in that, include: The update status of the lawn health semantic map of the target lawn area is monitored. The lawn health semantic map divides the target lawn area into multiple grid cells, and each grid cell is identified in real time with lawn health status information and area location information. When the lawn health status information of a certain grid cell on the lawn health semantic map changes from normal to abnormal, a maintenance task is generated based on the new lawn health status information of the grid cell and the regional location information. According to the preset task-device mapping rules, the maintenance task is sent to the corresponding task execution device.

2. The method for autonomous lawn maintenance and management according to claim 1, characterized in that, Also includes: Obtain the lawn health status identification result in the target lawn area, wherein the lawn health status identification result includes the current lawn health status, the confidence level of the current identification result, the current area location information, and the current timestamp; Obtain the set of grid cells corresponding to the current area location information from the lawn health semantic map; Based on the lawn health status recognition results, the layer information of each grid cell in the grid cell set is updated according to the preset data update rules to obtain the updated lawn health semantic map.

3. The method for autonomous lawn maintenance and management according to claim 2, characterized in that, The step of obtaining the lawn health status identification result in the target lawn area includes: Obtain raw sensing data of the target lawn area, wherein the raw sensing data includes at least one of RGB image blocks, elevation blocks, vegetation index blocks, and real-time vibration data; Based on the original sensing data, the lawn health status recognition result of the target lawn area is obtained through a preset lawn health status recognition model.

4. The method for autonomous lawn maintenance and management according to claim 3, characterized in that, Before acquiring the raw sensory data of the region to be tested, the method further includes: Construct the preset lawn health status recognition model; Specifically, historical RGB image patches, historical elevation patches, historical vegetation index patches, historical vibration data, and health status labels of the lawn area are obtained. The health status labels include coarse-grained health status labels and problem type labels. Based on the historical RGB image patches, historical elevation patches, historical vegetation index patches, historical vibration data, and the health status labels, a multi-branch neural network architecture is used, employing a progressive training strategy and an adversarial mode dropout training method to train the model to learn the feature changes of the historical RGB image patches, historical elevation patches, historical vegetation index patches, and historical vibration data under different health states, thereby obtaining the lawn health status recognition model.

5. The method for autonomous lawn maintenance and management according to any one of claims 1-4, characterized in that, The preset data update rules include: If the original lawn health status is healthy or unknown, or if the difference between the confidence level of the current lawn health status identification result and the confidence level of the original lawn health status identification result is greater than or equal to the first preset value, the original lawn health status identification result is updated with the current lawn health status identification result. If the current lawn health status is the same as the original lawn health status, update the timestamp; If the current lawn health status is different from the original lawn health status, and the difference between the confidence level of the current lawn health status identification result and the confidence level of the original lawn health status identification result is less than a second preset value, a conflict negotiation mechanism is triggered. The conflict negotiation mechanism includes any one of the following or any combination thereof: retaining the lawn health status label with higher confidence, marking the lawn health status label as pending confirmation, or making a decision based on the logical priority of the problem type.

6. The method for autonomous lawn maintenance and management according to any one of claims 1-4, characterized in that, Also includes: Obtain the status and location information of the task execution device; Based on the status information and the location information, combined with the maintenance task and the preset task priority sorting rules, an optimal task sequence and an optimal movement path are generated for the task execution device.

7. The method for autonomous lawn maintenance and management according to any one of claims 1-4, characterized in that, Also includes: Receive task completion information from the task execution device and mark the maintenance task as pending verification on the lawn health semantic map; The update status of the lawn health semantic map is continuously monitored. When the lawn health status information of the area corresponding to the maintenance task changes from abnormal to normal, the maintenance task is determined to be executed successfully. When the lawn health status information of the area corresponding to the maintenance task remains abnormal, the maintenance task is determined to be executed unsuccessfully. If the maintenance task is determined to have failed, the maintenance task will be regenerated or an alarm will be triggered.

8. A self-maintenance and management device for lawns, characterized in that, include: The monitoring module is used to monitor the update status of the lawn health semantic map of the target lawn area. The lawn health semantic map divides the target lawn area into multiple grid cells, and each grid cell is identified in real time with lawn health status information and area location information. The task generation module is used to generate a maintenance task based on the new lawn health status information and regional location information of the grid cell when the lawn health status information of a certain grid cell on the lawn health semantic map changes from normal to abnormal. The task sending module is used to send maintenance tasks to the corresponding task execution devices according to preset task-device mapping rules.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the lawn autonomous maintenance and management method as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the lawn autonomous maintenance and management method as described in any one of claims 1 to 7.

11. A computer program product comprising a computer program that can be stored on a non-transitory computer-readable storage medium, characterized in that, When the computer program is executed by the processor, the computer is able to perform the lawn autonomous maintenance and management method according to any one of claims 1 to 7.

12. A lawn autonomous maintenance management system for implementing the lawn autonomous maintenance management method according to any one of claims 1 to 7, characterized in that, The system includes inspection equipment, a server, and task execution equipment. The inspection equipment is used to inspect the target lawn area and update the lawn health status information and area location information of each grid cell on the lawn health semantic map in real time. The server is used to monitor the update status of the lawn health semantic map, generate maintenance tasks, and send the maintenance tasks to the corresponding task execution equipment. The task execution equipment is used to receive the maintenance tasks sent by the server and execute them.

Images