Unmanned agricultural production system based on agricultural scene and inspection robot control method

Abstract

The invention discloses an unmanned agricultural production system based on an agricultural scene and an inspection robot control method. The method comprises the steps that an environment sensing module obtains environment data in the agricultural scene through a plurality of sensors; the data processing module analyzes the state of the crops in the environment data and formulates a plurality of farming tasks according to the state of the crops; a task planning module formulates an inspection robot control strategy according to the priorities of the farming tasks and the state of the inspection robot; the robot control module controls movement of the inspection robot according to the inspection robot control strategy and completes corresponding farming tasks in the moving process, and unmanned production of a farm is achieved; according to the invention, corresponding farming tasks can be formulated according to crop states, the optimal moving path and task execution sequence of the robot in the farm are planned according to the priorities of the farming tasks, the full-range patrol inspection of the farm is realized, the production efficiency is improved, and the unmanned and intelligent farm is realized.

Description

Technical Field

[0001] The present invention relates to the field of unmanned production technology in farms, and in particular to an unmanned agricultural production system based on agricultural scenarios and a patrol robot control method. Background Art

[0002] Traditional agricultural production methods rely on a large amount of manpower to carry out farming operations, which is not only labor-intensive but also inefficient, affecting the output of agricultural products; traditional farming operations rely on experience and fixed cycles, which are difficult to meet the precise management of different crops; some existing agricultural production methods monitor the farm environment by setting sensors, but the collected information is relatively scattered, which is difficult to accurately reflect the growth status of crops, and will affect the decision-making of farming tasks; the operation mode of existing farm robots is mostly based on fixed rules, and it is impossible to adjust the operation status according to changes in crop status and the real-time status of the robot, which will affect the farming effect and resource utilization efficiency.

[0003] For example, a Chinese patent with authorization announcement number CN112790043B discloses a smart farm management system, including a control center, several ecological smart trees, a greenhouse breeding center, an outdoor automatic sprinkler unit, an outdoor plant light unit, a fully automatic slide-rail irrigation unit and a captive breeding center that are communicatively connected to the control center. The control center obtains real-time monitoring information from the ecological smart trees, the greenhouse breeding center, the outdoor automatic sprinkler unit, the outdoor plant light unit, the fully automatic slide-rail irrigation unit and the captive breeding center. In actual application, due to the presence of several ecological smart trees, greenhouse breeding centers, outdoor automatic sprinkler units, outdoor plant light units, fully automatic slide-rail irrigation units and captive breeding centers that are communicatively connected to the control center, and the deployment of several different types of sensor components, a lot of monitoring information of the smart farm can be obtained in real time, and timely feedback can be given to the control center for analysis and processing, and information can also be shared with remote mobile terminals to improve the level of intelligence.

[0004] The above-mentioned prior arts have the problems raised by this background technology: the information collected by sensors in some existing agricultural production is relatively scattered, which is difficult to accurately reflect the growth status of crops, and will affect the decision-making of agricultural tasks; the operating status of existing farm robots cannot be adjusted according to changes in crop status and the real-time status of the robot, which will affect agricultural results and resource utilization efficiency; in order to solve the above problems, the present invention proposes an unmanned agricultural production system based on agricultural scenarios and a patrol robot control method. Summary of the invention

[0005] In view of the shortcomings of the prior art, the main purpose of the present invention is to provide an unmanned agricultural production system and a patrol robot control method based on agricultural scenarios, which can effectively solve the problems in the background technology. The specific technical solutions of the present invention are as follows:

[0006] Unmanned agricultural production system based on agricultural scenarios, including:

[0007] Environmental perception module, data processing module, task planning module and robot control module; among which:

[0008] The environmental perception module acquires environmental data in the agricultural scene through multiple sensors;

[0009] The data processing module analyzes the state of the crops in the environmental data and formulates a plurality of farming tasks according to the state of the crops;

[0010] The task planning module formulates a control strategy for the inspection robot according to the priorities of the multiple farming tasks and the status of the inspection robot;

[0011] The robot control module controls the movement of the inspection robot according to the inspection robot control strategy, and completes corresponding farming tasks during the movement.

[0012] Specifically, the environment perception module includes:

[0013] An image acquisition unit and a soil moisture sensor unit; wherein:

[0014] The image acquisition unit collects image data of the farm through a camera;

[0015] The soil moisture sensor unit collects soil moisture data of a farm through a moisture sensor.

[0016] Specifically, the data processing module includes:

[0017] Data receiving unit, data analysis unit and task formulation unit; wherein:

[0018] The data receiving unit acquires image data and soil moisture data of the farm;

[0019] The data analysis unit analyzes the image data and soil moisture data to obtain the crop status in each grid in the farm;

[0020] The task formulation unit formulates agricultural tasks for the corresponding grid according to the crop status in each grid, wherein the agricultural tasks include watering tasks, fertilizing tasks and pesticide application tasks.

[0021] Specifically, the task planning module includes:

[0022] Task evaluation unit, robot state detection unit and path planning unit; wherein:

[0023] The task evaluation unit uses a preset task priority evaluation model to calculate the priority of each grid farming task in the farm, and sorts the tasks from high to low according to the priority to obtain the farming task sequence;

[0024] The robot state detection unit detects the state of the inspection robot in real time to obtain the state of the inspection robot;

[0025] The path planning unit formulates the inspection trajectory of the inspection robot by using the path planning method according to the agricultural task sequence and the state of the inspection robot, and obtains the control strategy of the inspection robot.

[0026] Specifically, the robot control module includes:

[0027] A motion control unit, an irrigation control unit, a fertilization control unit and a pesticide application control unit; wherein:

[0028] The motion control unit controls the movement of the inspection robot according to the inspection trajectory;

[0029] The irrigation control unit controls the inspection robot to perform irrigation according to the irrigation task in the farming task;

[0030] The fertilization control unit controls the inspection robot to fertilize according to the fertilization task in the farming task;

[0031] The pesticide application control unit controls the inspection robot to apply pesticides according to the pesticide application tasks in the farming tasks.

[0032] A patrol robot control method is used to control the patrol robot in the unmanned agricultural production system based on agricultural scenes, comprising:

[0033] Divide the agricultural scene into grids and obtain the environmental data of each grid in the agricultural scene through multiple sensors;

[0034] According to the environmental data, the state of the crops in each grid is analyzed by using a state analysis method to obtain the state of the crops;

[0035] According to the crop status, formulate farming tasks for each grid;

[0036] Using the preset task priority evaluation model, the priority of each grid farming task is calculated, and the tasks are sorted from high to low in order to obtain the order of farming tasks.

[0037] Formulate a control strategy for the inspection robot according to the agricultural task sequence and the status of the inspection robot;

[0038] According to the inspection robot control strategy, the movement of the inspection robot is controlled, and the corresponding agricultural tasks are completed during the movement.

[0039] Specifically, the state of the crops in each grid is analyzed using a state analysis method according to the environmental data to obtain the crop state, including:

[0040] According to the farm image data in the environmental data, the crop type in each grid is obtained using a preset crop type recognition model;

[0041] According to the crop types and crop image data, using a preset growth condition recognition model, the growth condition of the crops in each grid is obtained;

[0042] According to the crop image data, the preset pest and disease identification model is used to obtain the crop pest and disease situation in each grid.

[0043] Specifically, the preset task priority evaluation model is used to calculate the priority of each grid farming task, and the farming task sequence is obtained by sorting the tasks from high to low according to the priority, including:

[0044] By calculating the distance similarity and farming task similarity of each grid, multiple grids are clustered using clustering method to obtain multiple grid blocks;

[0045] According to the preset task priority evaluation model, the priority order of the agricultural tasks of the grid blocks is calculated to obtain the global agricultural task order;

[0046] In each grid block, the priority of the agricultural tasks of each grid is calculated through the preset task priority evaluation model to obtain the local agricultural task order.

[0047] Specifically, formulating a control strategy for the inspection robot according to the farming task sequence and the state of the inspection robot includes:

[0048] According to the global farming task sequence, the A* method is used to avoid farm environment obstacles, and the movement path of the inspection robot is globally planned to obtain a global movement path;

[0049] According to the local farming task sequence, in each grid block, the global moving path is locally planned using a dynamic window method to obtain a local moving path;

[0050] According to the state of the inspection robot and the agricultural tasks in the moving path, the local moving path is optimized to obtain the inspection robot control strategy.

[0051] Specifically, the local moving path is optimized according to the state of the inspection robot and the agricultural tasks in the moving path to obtain the inspection robot control strategy, including:

[0052] Monitor the remaining power, fertilizer and medicine amount of the inspection robot in real time to obtain the status of the inspection robot;

[0053] Generate the charging, feeding or medicine adding tasks of the inspection robot according to the inspection robot status and the order of agricultural tasks in the moving path;

[0054] Based on the charging, feeding or dosing tasks of the inspection robot, the local moving path is adjusted and optimized using the preset path optimization model to obtain the inspection robot control strategy.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] The present invention collects farm environmental data through the environmental perception module and provides environmental data support; through the data processing module, it identifies the state of crops and formulates corresponding farming tasks, accelerates the growth cycle of crops, and improves the yield and quality of crops; through the task planning module, it plans the optimal path for the robot to move in the farm and the task execution sequence, optimizes the working process of the inspection robot, avoids invalid paths and repetitive labor, and realizes inspection of the entire farm, ensures the continuity of farm production, and improves production efficiency; through the robot control module, it controls the robot to move according to the robot control strategy and executes corresponding farming tasks, realizes remote and precise control of the inspection robot, and thus realizes unmanned, intelligent and continuous production of the farm. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 This is a schematic diagram of the structure of an unmanned agricultural production system based on an agricultural scenario in Example 1 of the present invention;

[0058] Figure 2 This is a workflow diagram of the inspection robot control method in Embodiment 2 of the present invention;

[0059] Figure 3 A schematic diagram of prioritizing farm grid tasks in Embodiment 2 of the present invention;

[0060] Figure 4 This is a workflow diagram for formulating a control strategy for the inspection robot in Example 2 of the present invention. DETAILED DESCRIPTION

[0061] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0062] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0063] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0064] Embodiment 1:

[0065] This embodiment provides an unmanned agricultural production system based on agricultural scenarios, such as Figure 1 As shown, the unmanned agricultural production system based on agricultural scenarios includes:

[0066] Environmental perception module, data processing module, task planning module and robot control module; among which:

[0067] The environmental perception module acquires environmental data in the agricultural scene through multiple sensors;

[0068] The data processing module analyzes the state of the crops in the environmental data and formulates a plurality of farming tasks according to the state of the crops;

[0069] The task planning module formulates a control strategy for the inspection robot according to the priorities of the multiple farming tasks and the status of the inspection robot;

[0070] The robot control module controls the movement of the inspection robot according to the inspection robot control strategy, and completes corresponding farming tasks during the movement to achieve unmanned production on the farm.

[0071] Since traditional agricultural production methods are time-consuming and labor-intensive, and it is difficult to accurately and timely monitor and regulate the crop growth environment, the yield and quality of crops are limited. The present invention proposes an unmanned agricultural production system based on agricultural scenarios to improve the intelligence and unmanned nature of the agricultural production process. By controlling the movement of the inspection robot, the full-range inspection of the farm is realized, the growth environment of crops is adjusted in time, and the yield and quality of crops are improved.

[0072] In this embodiment, an unmanned agricultural production system based on an agricultural scenario includes an environmental perception module, a data processing module, a task planning module and a robot control module. First, the environmental perception module collects environmental data in an all-round and real-time manner, including information such as soil moisture, crop growth form and signs of pests and diseases, to provide data support for subsequent decision-making; the data processing module uses an image recognition method to process and analyze the environmental data to obtain the growth and health status of crops, and then formulates a series of agricultural tasks that are adapted to current needs, such as irrigation, fertilization, and pest control tasks; the task planning module uses an intelligent optimization algorithm to plan a most efficient and reasonable patrol robot action route and task execution sequence based on the agricultural tasks in different areas of the farm, combined with the real-time status of the inspection robot in the farm, including power, fertilizer, and medicine, as well as the priority determined by the urgency and importance of each task, to form a patrol robot control strategy; finally, the robot control module controls the robot to move along the planned path according to the patrol robot control strategy, and at each node of the movement, drives the robot to complete the corresponding agricultural tasks, such as turning on the irrigation nozzle, releasing pesticides, etc., to realize unmanned and intelligent continuous production of the farm.

[0073] Specifically, the environmental perception module collects farm environmental data through sensors, provides comprehensive and real-time agricultural scene information, and provides data support for subsequent decision-making; the data processing module processes the farm environmental data, identifies the status of crops and formulates corresponding agricultural tasks, converts complex environmental data into operational agricultural instructions, and provides support for precision agriculture; the task planning module plans the optimal path and task execution sequence for the inspection robot to move within the farm according to the agricultural task priority and the status of the inspection robot, which can optimize the working process of the inspection robot, avoid invalid paths and repetitive work, and realize full-range inspection of the farm, ensuring the continuity of farm production and improving production efficiency; the robot control module controls the movement of the robot and performs corresponding agricultural tasks according to the robot control strategy, realizes remote and precise control of the inspection robot, and ensures the continuous work of the robot.

[0074] The present invention collects farm environmental data through the environmental perception module and provides environmental data support; through the data processing module, it identifies the state of crops and formulates corresponding farming tasks, accelerates the growth cycle of crops, and improves the yield and quality of crops; through the task planning module, it plans the optimal path for the robot to move in the farm and the task execution sequence, optimizes the working process of the inspection robot, avoids invalid paths and repetitive labor, and realizes inspection of the entire farm, ensures the continuity of farm production, and improves production efficiency; through the robot control module, it controls the robot to move according to the robot control strategy and executes corresponding farming tasks, realizes remote and precise control of the inspection robot, and thus realizes unmanned, intelligent and continuous production of the farm.

[0075] Furthermore, the environment perception module includes:

[0076] An image acquisition unit and a soil moisture sensor unit; wherein:

[0077] The image acquisition unit collects image data of the farm through a camera;

[0078] The soil moisture sensor unit collects soil moisture data of a farm through a moisture sensor.

[0079] In this embodiment, the environmental perception module includes an image acquisition unit and a soil moisture sensor unit, wherein the image acquisition unit collects image data of crops in the farm through a camera. Based on the image data, an image recognition method can be used to identify abnormal conditions such as yellowing, withering of crop leaves, and holes formed by pests, and formulate corresponding farming tasks based on the abnormal conditions; the soil moisture sensor unit obtains the soil moisture conditions in the farm through a humidity sensor, and formulates corresponding irrigation tasks based on the soil moisture. By grasping the soil moisture conditions in real time, it is possible to avoid blind irrigation that causes waste of water resources or affects crop growth due to water shortage, thereby ensuring that the soil moisture is always maintained in a range suitable for crop growth, thereby achieving water saving and efficiency improvement.

[0080] Furthermore, the data processing module includes:

[0081] Data receiving unit, data analysis unit and task formulation unit; wherein:

[0082] The data receiving unit acquires image data and soil moisture data of the farm;

[0083] The data analysis unit analyzes the image data and soil moisture data to obtain the crop status in each grid in the farm;

[0084] The task formulation unit formulates agricultural tasks for the corresponding grid according to the crop status in each grid, wherein the agricultural tasks include watering tasks, fertilizing tasks and pesticide application tasks.

[0085] In this embodiment, the data processing module includes a data receiving unit, a data analysis unit and a task formulation unit, wherein the data receiving unit obtains the farm image data and soil moisture data collected by the environmental perception module, and provides a data basis for data analysis by obtaining the data; the data analysis unit uses an image recognition method to identify the crop type, pest and disease situation, growth stage and other crop condition information for the received image data, and for the soil moisture data, combines the farm soil type and the water requirement characteristics of different crops to judge the degree of soil dryness and wetness, and on this basis, divides the farm into pre-divided grid areas as units, and judges the status of the crops in each grid, including health status, growth progress and water supply and demand.

[0086] Specifically, the task formulation unit has a complete agricultural knowledge base built in, including agricultural demand standards for different crops at various growth stages and in different health conditions. After receiving the status of crops in each grid from the data analysis unit, intelligent matching is performed according to the knowledge base rules. For example, when the soil moisture of crops is low and they are in a period of vigorous growth, irrigation tasks are formulated to determine the irrigation time and amount of water. Corresponding agricultural tasks are formulated for the crops in each grid. When there are no abnormal conditions for the crops in the grid, the agricultural tasks of the grid are set as inspection tasks. Automatic formulation of agricultural tasks based on the status of crops can realize automated agricultural planning, greatly shorten the time from discovering crop problems to formulating response measures, ensure timely implementation of agricultural activities, and improve agricultural production efficiency. Through the automatic formulation of agricultural tasks, the standardization of agricultural operations is strengthened, which can avoid the arbitrariness and non-standardization of human operations, thereby improving the stability of crop quality.

[0087] Furthermore, the task planning module includes:

[0088] Task evaluation unit, robot state detection unit and path planning unit; wherein:

[0089] The task evaluation unit uses a preset task priority evaluation model to calculate the priority of each grid farming task in the farm, and sorts the tasks from high to low according to the priority to obtain the farming task sequence;

[0090] The robot state detection unit detects the state of the inspection robot in real time to obtain the state of the inspection robot;

[0091] The path planning unit formulates the inspection trajectory of the inspection robot by using the path planning method according to the agricultural task sequence and the state of the inspection robot, and obtains the control strategy of the inspection robot.

[0092] In this embodiment, the task planning module includes a task evaluation unit, a robot state detection unit and a path planning unit, wherein the task evaluation unit sets the priority of the task according to a preset task priority evaluation model, combined with the growth conditions, water shortage conditions and pest and disease conditions of the crops. For example, when the severity of the pests and diseases of the crops is high, the corresponding pesticide application task is set with a higher priority, and urgent tasks are given priority. The priority value of the agricultural tasks of each grid in the farm is calculated, and they are sorted in order from large to small to obtain the order of agricultural tasks to be executed, so as to provide task execution sequence guidance for the subsequent path planning of the inspection robot. By prioritizing the agricultural tasks, the most urgent tasks that have the greatest impact on the growth of crops can be executed first, thereby ensuring the smooth growth process of the crops.

[0093] Specifically, the robot status detection unit performs real-time detection of the status of the inspection robot, including the robot's power, position coordinates, remaining fertilizer, remaining medicine and other status information. By detecting the robot's status in real time, when the robot is low on power or fertilizer, the robot can be promptly controlled to return to the power station for charging or replenishing materials in time, thus avoiding the robot from shutting down on the way and providing continuity for the robot's operation. It can also arrange nearby and appropriate agricultural tasks according to the robot's position and status, thus reducing the robot's empty running distance and improving overall operation efficiency.

[0094] Specifically, the path planning unit divides the farm area into multiple grids according to the order of agricultural tasks and the status of the inspection robot, analyzes the priority of agricultural tasks corresponding to each grid, sets the order of grids to be reached, and uses the path planning method to take the current position of the robot as the starting point, continuously searches for feasible nodes, and screens out an optimal moving path by calculating the comprehensive cost of different paths, including path length, task execution efficiency and other costs. Finally, a patrol robot control strategy containing detailed instruction sequences such as speed, steering, and operation start and stop is generated to guide the robot to complete inspection and agricultural tasks efficiently. By planning the mobile path of the patrol robot, the robot can complete agricultural tasks with the shortest path and the fastest speed, reducing invalid movement time, and significantly improving farm production efficiency. When faced with sudden situations such as temporary obstacles on the farm and new emergency tasks, the path planning method can quickly re-plan the path to ensure the stable operation of the robot, adapt to complex environmental requirements, and ensure orderly agricultural production.

[0095] Furthermore, the robot control module comprises:

[0096] A motion control unit, an irrigation control unit, a fertilization control unit and a pesticide application control unit; wherein:

[0097] The motion control unit controls the movement of the inspection robot according to the inspection trajectory;

[0098] The irrigation control unit controls the inspection robot to perform irrigation according to the irrigation task in the farming task;

[0099] The fertilization control unit controls the inspection robot to fertilize according to the fertilization task in the farming task;

[0100] The pesticide application control unit controls the inspection robot to apply pesticides according to the pesticide application tasks in the farming tasks.

[0101] In this embodiment, the robot control module includes a motion control unit, an irrigation control unit, a fertilization control unit and a pesticide control unit, wherein the motion control unit controls the robot to move smoothly and efficiently along a preset inspection trajectory according to the robot control strategy generated by the path planning unit, and accurately stops at each agricultural task operation point. By controlling the robot to move strictly along the planned path, it can avoid deviations from the route and other problems such as trampling on farmland and missing tasks, thereby improving the accuracy of the operation. After the irrigation control unit obtains the irrigation task instructions in the agricultural task, the inspection robot turns on or off the water supply switch according to the parameters such as the irrigation water volume and irrigation time in the task instructions, so as to achieve accurate and uniform irrigation of the designated farmland area and avoid waste of water resources. By irrigating crops at a fixed point and in a fixed quantity, it can meet the growth needs and prevent over-irrigation, save water costs, operate at night or during periods when manpower is inconvenient, improve the timeliness of irrigation, ensure that the growth of crops is not affected by water shortage, and achieve automated operation.

[0102] Specifically, after receiving the fertilization task, the fertilization control unit controls the operation of the robot's fertilization device according to parameters such as the amount of fertilizer and the type of fertilizer, performs fertilization operations, evenly spreads fertilizers, and meets the nutrient needs of crops. Through precise fertilization, fertilizer cost expenditures can be reduced to prevent uneven growth of crops due to uneven fertilization. By automatically executing the fertilization process and cooperating with the robot's inspection route to efficiently complete the fertilization task, the efficiency and scientificity of fertilization operations can be improved. After receiving the pesticide task, the pesticide control unit controls the dosing device according to parameters such as the type of pesticide and spraying dosage, mixes and sprays the pesticide according to the amount of medicine and water, and achieves accurate and comprehensive pesticide application in the pest-affected area. By applying pesticides according to the needs of pest and disease control, the prevention and control effect can be ensured and the abuse of pesticides can be avoided. By quickly responding to pesticide tasks and preventing and controlling pests and diseases in a timely manner, the risk resistance of the agricultural production system can be improved.

[0103] Embodiment 2:

[0104] This embodiment provides a patrol robot control method for controlling the patrol robot in the unmanned agricultural production system based on the agricultural scene, such as Figure 2 As shown, the inspection robot control method comprises:

[0105] S101, dividing the agricultural scene into grids, and obtaining environmental data of each grid in the agricultural scene through multiple sensors;

[0106] S102, analyzing the state of the crops in each grid using a state analysis method according to the environmental data to obtain the crop state;

[0107] S103, formulating farming tasks for each grid according to the crop status;

[0108] S104, using a preset task priority evaluation model, calculating the priority of each grid farming task, and sorting them from high to low according to the priority to obtain the farming task sequence;

[0109] S105, formulating a control strategy for the inspection robot according to the agricultural task sequence and the state of the inspection robot;

[0110] S106. According to the inspection robot control strategy, the movement of the inspection robot is controlled, and corresponding farming tasks are completed during the movement to realize unmanned production on the farm.

[0111] In this embodiment, the inspection robot control method is used to control the movement and working state of the inspection robot in the farm in real time to complete the corresponding agricultural tasks. First, the farm environment is divided into grids, and the farm environment data of each grid collected by the environmental perception module is obtained. Through grid division, complex farms can be managed in a grid manner, which is convenient for accurately locating problem areas and providing space for agricultural task decision-making; according to real-time environmental data, combined with historical environmental data and crop growth records in the database, the crop image data is analyzed, and the crop growth, pest and disease, water shortage and other crop conditions are obtained. For example, through image recognition, it is found that there are bacterial spots or wormholes on the leaves of crops, indicating that there are pests and diseases in the crop area; through image analysis, abnormal conditions of crops can be discovered in time, and corresponding strategies can be formulated to improve the quality and yield of crops.

[0112] Specifically, according to the growth status of crops in each grid, corresponding agricultural tasks are automatically formulated in combination with the agricultural knowledge base. For example, when pests and diseases are found, the corresponding pesticides are selected with reference to the knowledge base, and the spraying concentration and dosage are determined. According to the crop status, the corresponding agricultural tasks are intelligently matched to ensure that the tasks of each grid are reasonably planned and the matching degree of agricultural tasks is improved. According to the urgency and importance of the agricultural tasks of each grid, the priority of agricultural tasks is scored based on a variety of factors to obtain the priority score of the agricultural tasks of each grid. The tasks are sorted from large to small according to the scores to obtain the order of agricultural tasks. By prioritizing agricultural tasks, urgent tasks can be given priority to meet agricultural production needs and avoid affecting crop growth, thereby effectively increasing crop yields.

[0113] Specifically, the robot's moving path is planned in combination with the order of agricultural tasks and the real-time status of the inspection robot. The factors such as farm terrain and environmental obstacles are considered, and the travel cost is set for each grid. The current position of the robot is taken as the starting point, and the corresponding target grid sequence is determined according to the order of agricultural tasks. The path planning method is used to search for the optimal path, and the path is optimized according to the state constraints such as the robot's power, remaining fertilizer and medicine amount, and the robot control strategy is generated. By planning and optimizing the robot's driving path, the robot's task execution efficiency can be improved. The driving path is optimized in real time according to the robot's status, which can flexibly respond to situations such as insufficient power and ensure the continuity of the robot's work; the robot's operation is driven according to the robot control strategy, and the control strategy is decomposed into execution instructions of each unit. The robot moves according to the instructions, slows down and stops accurately when approaching the operation point, and the corresponding control unit starts the operation equipment, monitors the task progress in real time, and moves to the next target grid after completing the current grid agricultural task. The robot continues to work. Through the precise control of the inspection robot, the robot can work uninterruptedly, ensuring that the state of crops in the farm grows according to the preset growth rate, improving production efficiency, and realizing unmanned and intelligent farms.

[0114] Furthermore, the state of the crops in each grid is analyzed using a state analysis method according to the environmental data to obtain the crop state, including:

[0115] S201, obtaining the crop type in each grid using a preset crop type recognition model according to the farm image data in the environmental data;

[0116] S202, obtaining the growth condition of the crops in each grid using a preset growth condition recognition model according to the crop types and crop image data;

[0117] S203: According to the crop image data, using a preset pest and disease identification model, obtain the crop pest and disease situation in each grid.

[0118] In this embodiment, since different types of crops have different growth requirements, first, based on the collected farm image data, a pre-trained crop type recognition model is used. Specifically, the model is a convolutional neural network model. The model can automatically extract feature information in the image, such as leaf shape, texture, color distribution and other feature information, and determine the type of crop in each grid based on the features. By automatically identifying the type of crop, the time and labor cost of manual identification can be saved, which is convenient for setting different farming tasks for different crops and improving the efficiency of agricultural production management. Specifically, based on the crop type and crop image data, a pre-trained growth condition recognition model is used, and a deep learning model is also used to extract corresponding features from the image data, including plant height, number of leaves, leaf size and other feature information. Based on the standard image features of crops at various growth stages, it is determined whether there are any abnormalities in the current growth of crops, such as growth retardation, stunted development, etc. Corresponding measures can be formulated based on the abnormalities. By performing real-time detection of the growth of crops, it is possible to understand what growth stage the crops are in, which helps to adjust farming tasks in a timely manner and ensure the normal growth of crops.

[0119] Specifically, the pre-trained pest and disease recognition model is used to identify crop image data. The type of pest and disease is determined based on the appearance characteristics of the pests and diseases in the image, such as the morphology, color, size of the pests, the shape, color, and distribution area of ​​the fungus after infection with the bacteria, and other characteristic information. The severity of the pests and diseases is estimated based on factors such as the proportion of the pest and disease area and the degree of feature prominence. The corresponding pesticide application tasks are formulated according to the pest and disease situation. By automatically identifying the pest and disease situation of crops, signs of crop pests and diseases can be discovered in a timely manner, the type and severity of pests and diseases can be clarified, and corresponding pesticide application strategies can be formulated to avoid blind use of pesticides, reduce pesticide waste and pesticide residues in agricultural products, and ensure the quality and safety of agricultural products.

[0120] Furthermore, the preset task priority evaluation model is used to calculate the priority of each grid farming task, and the farming task sequence is obtained by sorting the tasks from high to low according to the priority, including:

[0121] S301, by calculating the distance similarity and farming task similarity of each grid, clustering multiple grids using a clustering method to obtain multiple grid blocks;

[0122] S302, calculating the priority order of agricultural tasks of the grid blocks according to a preset task priority evaluation model, and obtaining a global agricultural task order;

[0123] S303: In each grid block, the priority of the agricultural tasks of each grid is calculated by a preset task priority evaluation model to obtain a local agricultural task sequence.

[0124] In this embodiment, if Figure 3 This is a schematic diagram for prioritizing farm grid tasks. First, the distance similarity between grids is measured by calculating the distance between the grid center coordinates. According to the farming task type and the corresponding task parameters, when irrigation, fertilization and pesticide application are not required in the grid, the farming task of the grid is set as an inspection task, and the crop status in the grid is detected in real time. For example, when two grids both require irrigation tasks with similar water volumes and the types of pesticides required for pest control are also the same, the farming task similarity between the two grids is relatively high. Using a clustering algorithm, combined with distance similarity and farming task similarity, these grids with high similarity are clustered into one category to form multiple grid blocks, so that the grids in the same block are similar in geographical location and farming needs, which can reduce the complexity of task planning. For example, robots can work continuously in the same grid block, reducing transfer costs and improving work efficiency.

[0125] Specifically, the evaluation index system for determining task priority includes four factors: crop growth, health, task time, and soil environment. A factor quantification standard is set, and the quantitative values ​​of the four factors of crop growth, health, task time, and soil environment are obtained according to the crop status. A weight is assigned to each factor. In this embodiment, the weight of crop growth is 0.4, the weight of health is 0.3, the weight of task time is 0.1, and the weight of soil environment is 0.2. According to the evaluation index, the task priority evaluation model is determined. The specific formula is as follows:

[0126]

[0127] In the formula, is the task priority value of the i-th grid, is the crop growth value of the i-th grid, is the crop health value of the i-th grid, is the agricultural task time value of the i-th grid, is the soil environment condition value of the ith grid. According to the task priority value of each grid, the task priority value of each grid block is calculated. The specific formula is as follows:

[0128]

[0129] In the formula, is the task priority value of the zth grid block, and r is the number of grids in the zth grid block. The agricultural task priority value of each grid block is calculated according to the formula. According to the agricultural task priority value, the grid blocks are sorted to obtain the global agricultural task order. By sorting the tasks globally, the grid block areas with urgent tasks can be processed first, avoiding excessive resource occupation by local tasks and ensuring overall production.

[0130] After calculating the global agricultural task sequence, within each grid block, the grids are prioritized according to the task priority value of each grid to obtain the local agricultural task sequence. By sorting each grid within the grid block, it is ensured that the special needs of crops in different grids within the same grid block are responded to in a timely manner, the operation accuracy is improved, and the quality of agricultural products is effectively improved.

[0131] Further, such as Figure 4 As shown, the control strategy of the inspection robot is formulated according to the farming task sequence and the state of the inspection robot, including:

[0132] S401, according to the global farming task sequence, using the A* method to avoid farm environment obstacles, and globally planning the movement path of the inspection robot to obtain a global movement path;

[0133] S402, according to the local farming task sequence, in each grid block, using a dynamic window method to perform local planning on the global moving path to obtain a local moving path;

[0134] S403: Optimize the local moving path according to the state of the inspection robot and the farming tasks in the moving path to obtain a control strategy for the inspection robot.

[0135] In this embodiment, first, according to the global farming task sequence, each grid block is used as a target node, and a moving path that can avoid obstacles in the farm environment and reach each grid block is planned through the A* method to achieve full-range inspection of the farm. Specifically, the center position of the target grid block is determined according to the global farming task sequence, and the corresponding target node sequence is generated. The current position of the robot is used as the starting point and the next grid block position is used as the end point. According to the radar configured by the robot, the environmental obstacles around the robot are detected in real time, and the detected obstacle coordinates are marked as inaccessible, ensuring that the searched path can avoid obstacles in the environment and ensure the stable operation of the robot.

[0136] Specifically, according to a preset evaluation function, the coordinate with the best evaluation result among the adjacent coordinates is selected as the next position coordinate. The evaluation function in this embodiment is:

[0137]

[0138] Where n is the next position coordinate in the path search process, is the evaluation function value, is the actual cost from the starting coordinate to the current coordinate, It is the estimated cost from the current coordinate to the end coordinate, and the Euclidean distance is used for evaluation. By calculating the evaluation function value of each adjacent coordinate, the coordinate with the smallest evaluation function value is selected as the next position coordinate. By selecting the position coordinate with the smallest evaluation function value, the shortest path can be gradually selected; the intermediate position coordinate sequence is obtained from the output of the A* algorithm, and these intermediate coordinates are connected, and gradually connected from the starting point to the end point to form the robot's moving path, ensuring that the object can move to the target grid block along a continuous, collision-free path, thereby improving the object's movement efficiency. Traverse and search according to the grid block target node sequence until the path of all grid block target nodes is found, and the global moving path is obtained. When one inspection is completed, the current position is used as the starting point, and the next inspection is carried out with the same workflow.

[0139] After calculating the global moving path, in each grid block area, according to each grid position corresponding to the local farming task sequence, a corresponding target grid point sequence is generated, and in each grid block, the farming tasks corresponding to each grid are completed in sequence. When the robot enters the grid block, it quickly obtains the current initial state information through its own configured sensors, including position coordinates, real-time speed, acceleration and other values, and calculates the robot's arrival area based on the robot's real-time motion state. The calculation formula is as follows:

[0140]

[0141] In the formula, is the time step The coordinates of the robot's arrival area. , are the horizontal and vertical coordinates of the robot’s current position, , The current speed and acceleration of the robot are respectively. According to the robot's performance parameters and task requirements, the speed window is initialized, and all speed combinations in the current speed window are traversed. For each speed combination, the arrival area of ​​the robot is predicted according to the arrival area calculation formula, and the arrival area is compared with the local obstacle position to determine whether a collision will occur. When it is determined that there is a risk of collision, the speed combination is immediately excluded; if there is no collision, the time required for the robot to reach the target point under the speed condition is calculated in detail, and the optimal speed combination that can reach the target grid point the fastest is selected, and the speed window is dynamically optimized and updated; according to the final optimal speed instruction sequence, the robot is driven to move steadily, and the position information of the robot at each time step is recorded in real time, and these position points are connected in sequence to finally form an accurate and smooth local moving path. In each grid block, each grid is traversed as the target grid point in order of priority until a local moving path to each grid point is obtained.

[0142] According to the robot's movement state, the robot's speed and acceleration are adjusted in real time to search for an optimal path to efficiently reach each grid, thereby improving the robot's work efficiency. It can also flexibly avoid sudden obstacles on the farm and accurately dock at the operating point of each grid, thereby improving the execution accuracy of agricultural tasks.

[0143] Specifically, the real-time status of the inspection robot is deeply integrated with the specific needs of the agricultural tasks that need to be performed on the mobile path, and the planned local mobile path is further optimized. For example, once the robot is detected to be low on power, the emergency mechanism is immediately activated, and a path segment leading to the charging area is planned first, and the execution order of subsequent agricultural tasks is reasonably adjusted simultaneously to ensure that the robot can replenish energy in time, ensure the continuity of the robot's work, and prevent downtime; for scenes that need to perform agricultural tasks such as irrigation, fertilization, and pesticide application, the system will reasonably allocate the dwell time between each operation point based on the time required for the task volume, to ensure that the robot maintains an efficient and stable operating state while completing the task. Finally, the all-round information such as the optimized path, speed, and operation start and stop is integrated to obtain the inspection robot control strategy.

[0144] Specifically, by optimizing the mobile path, the task schedule and path planning can be intelligently adjusted according to the real-time status information of the robot to ensure that the robot can operate continuously and stably, effectively avoiding production interruptions caused by emergencies such as power exhaustion and insufficient fertilizer, and ensuring the continuity of agricultural production.

[0145] Furthermore, the local moving path is optimized according to the state of the inspection robot and the agricultural tasks in the moving path to obtain the inspection robot control strategy, including:

[0146] S501, real-time monitoring of the remaining power, fertilizer amount, and medicine amount of the inspection robot to obtain the status of the inspection robot;

[0147] S502, generating charging, feeding or adding medicine tasks for the inspection robot according to the state of the inspection robot and the order of agricultural tasks in the moving path;

[0148] S503: Based on the charging, feeding or dosing tasks of the inspection robot, the local moving path is adjusted and optimized using a preset path optimization model so that the moving path of the inspection robot can cover the entire farm, thereby obtaining a control strategy for the inspection robot.

[0149] In this embodiment, the real-time status of the robot is monitored by various sensors configured in the robot to obtain the remaining power, fertilizer amount and medicine amount of the inspection robot; according to the preset task triggering rules, the corresponding charging, feeding or medicine adding tasks are generated. The task triggering rules are: when When the charging task is triggered, is the remaining battery power of the robot, The power required for the robot to complete subsequent tasks, is the power threshold; when When , the feeding task is triggered, where, is the remaining fertilizer amount of the robot, The amount of fertilizer required for the robot's subsequent tasks; When , the dosing task is triggered, where, is the remaining pesticide amount of the robot, The amount of pesticide required for the robot's subsequent tasks. By predicting resource shortages in advance and automatically generating replenishment tasks, it is possible to avoid downtime or task interruptions during the robot's work and improve production efficiency.

[0150] Specifically, according to the charging, feeding or dosing tasks of the inspection robot, the optimal path is searched and a path optimization model is constructed under the premise of meeting the task sequence. The model formula is as follows:

[0151]

[0152]

[0153]

[0154]

[0155] In the formula, m is the total number of paths, is the weight of the j-th path, is the length of the j-th path, is the energy consumption weight, is the total number of supply points, is the resource supply of the tth supply point, is the time weight, is the total number of tasks, is the execution time of the kth task. The model aims to minimize the total distance, balance resource consumption, and complete the task on time. It uses the particle swarm optimization algorithm to iteratively search for the optimal path. When the number of iterations reaches the preset maximum number of iterations, the solution with the optimal value of the corresponding objective function is taken as the final solution to obtain the inspection robot control strategy. By adding corresponding charging, feeding or adding medicine tasks to the robot, the robot can efficiently switch between task execution and supply, improve the overall operation efficiency, ensure that the robot has sufficient resources, avoid difficulties due to lack of electricity, materials, and medicine, and ensure the continuous and stable advancement of agricultural production.

[0156] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. Unmanned agricultural production system based on agricultural scenarios, characterized by: include: Environmental perception module, data processing module, task planning module and robot control module; among which: The environmental perception module acquires environmental data in the agricultural scene through multiple sensors; The data processing module analyzes the state of the crops in the environmental data and formulates a plurality of farming tasks according to the state of the crops; The task planning module formulates a control strategy for the inspection robot according to the priorities of the multiple farming tasks and the status of the inspection robot; The robot control module controls the movement of the inspection robot according to the inspection robot control strategy, and completes corresponding farming tasks during the movement.

2. The unmanned agricultural production system based on agricultural scenes according to claim 1 is characterized in that: The environment perception module comprises: An image acquisition unit and a soil moisture sensor unit; wherein: The image acquisition unit collects image data of the farm through a camera; The soil moisture sensor unit collects soil moisture data of a farm through a moisture sensor.

3. The unmanned agricultural production system based on agricultural scenes according to claim 1 is characterized in that: The data processing module comprises: Data receiving unit, data analysis unit and task formulation unit; wherein: The data receiving unit acquires image data and soil moisture data of the farm; The data analysis unit analyzes the image data and soil moisture data to obtain the crop status in each grid in the farm; The task formulation unit formulates agricultural tasks for the corresponding grid according to the crop status in each grid, wherein the agricultural tasks include watering tasks, fertilizing tasks and pesticide application tasks.

4. The unmanned agricultural production system based on agricultural scenes according to claim 1 is characterized in that: The mission planning module comprises: Task evaluation unit, robot state detection unit and path planning unit; wherein: The task evaluation unit uses a preset task priority evaluation model to calculate the priority of each grid farming task in the farm, and sorts the tasks from high to low according to the priority to obtain the farming task sequence; The robot state detection unit detects the state of the inspection robot in real time to obtain the state of the inspection robot; The path planning unit formulates the inspection trajectory of the inspection robot by using the path planning method according to the agricultural task sequence and the state of the inspection robot, and obtains the control strategy of the inspection robot.

5. The unmanned agricultural production system based on agricultural scenes according to claim 1 is characterized in that: The robot control module comprises: A motion control unit, an irrigation control unit, a fertilization control unit and a pesticide application control unit; wherein: The motion control unit controls the movement of the inspection robot according to the inspection trajectory; The irrigation control unit controls the inspection robot to perform irrigation according to the irrigation task in the farming task; The fertilization control unit controls the inspection robot to fertilize according to the fertilization task in the farming task; The pesticide application control unit controls the inspection robot to apply pesticides according to the pesticide application tasks in the farming tasks.

6. A patrol robot control method, used to control the patrol robot in the unmanned agricultural production system based on agricultural scenes as described in any one of claims 1 to 5, characterized in that: include: Divide the agricultural scene into grids and obtain the environmental data of each grid in the agricultural scene through multiple sensors; According to the environmental data, the state of the crops in each grid is analyzed by using a state analysis method to obtain the state of the crops; According to the crop status, formulate farming tasks for each grid; Using the preset task priority evaluation model, the priority of each grid farming task is calculated, and the tasks are sorted from high to low in order to obtain the order of farming tasks. Formulate a control strategy for the inspection robot according to the agricultural task sequence and the status of the inspection robot; According to the inspection robot control strategy, the movement of the inspection robot is controlled, and the corresponding agricultural tasks are completed during the movement.

7. A patrol robot control method according to claim 6, characterized in that: The state analysis method is used to analyze the state of the crops in each grid according to the environmental data to obtain the state of the crops, including: According to the farm image data in the environmental data, the crop type in each grid is obtained using a preset crop type recognition model; According to the crop types and crop image data, using a preset growth condition recognition model, the growth condition of the crops in each grid is obtained; According to the crop image data, the preset pest and disease identification model is used to obtain the crop pest and disease situation in each grid.

8. The inspection robot control method according to claim 6, characterized in that: The preset task priority evaluation model is used to calculate the priority of each grid farming task, and the farming task sequence is obtained by sorting the tasks from high to low according to the priority, including: By calculating the distance similarity and farming task similarity of each grid, multiple grids are clustered using clustering method to obtain multiple grid blocks; According to the preset task priority evaluation model, the priority order of the agricultural tasks of the grid blocks is calculated to obtain the global agricultural task order; In each grid block, the priority of the agricultural tasks of each grid is calculated through the preset task priority evaluation model to obtain the local agricultural task order.

9. The inspection robot control method according to claim 8, characterized in that: The control strategy of the inspection robot is formulated according to the agricultural task sequence and the state of the inspection robot, including: According to the global farming task sequence, the A* method is used to avoid farm environment obstacles, and the movement path of the inspection robot is globally planned to obtain a global movement path; According to the local farming task sequence, in each grid block, the global moving path is locally planned using a dynamic window method to obtain a local moving path; According to the state of the inspection robot and the agricultural tasks in the moving path, the local moving path is optimized to obtain the inspection robot control strategy.

10. The inspection robot control method according to claim 9, characterized in that: The method of optimizing the local moving path according to the state of the inspection robot and the agricultural tasks in the moving path to obtain the inspection robot control strategy includes: Monitor the remaining power, fertilizer and medicine amount of the inspection robot in real time to obtain the status of the inspection robot; Generate the charging, feeding or medicine adding tasks of the inspection robot according to the inspection robot status and the order of agricultural tasks in the moving path; Based on the charging, feeding or dosing tasks of the inspection robot, the local moving path is adjusted and optimized using the preset path optimization model to obtain the inspection robot control strategy.

Images