Intelligent pesticide spraying path planning system and method

Abstract

The invention discloses an intelligent pesticide spraying path planning system and method, and relates to the technical field of pesticide spraying path planning, the system is composed of a plurality of function modules, and each function module operates in sequence, namely an information acquisition module, a data analysis processing module, a path planning module, a pesticide spraying adjustment module and an evaluation adjustment module; the technical key points are that the system integrates the functions of information acquisition, path planning, real-time identification and dynamic adjustment, realizes comprehensive intelligence of orchard spraying operation, improves the precision and efficiency of spraying operation, ensures the high efficiency and accuracy of spraying operation by automatically adjusting the estimated spraying flow and the opening number of each spraying head, and reduces the labor intensity of workers. Meanwhile, the fruit trees are fully covered with the pesticide, the pesticide spraying amount and the pesticide spraying flow are accurately controlled, excessive use and drifting of the pesticide are reduced, and pollution to the environment is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of spraying path planning, and in particular to an intelligent spraying path planning system and method. Background Art

[0002] Spraying path planning refers to the process of pre-planning the driving path of spraying equipment (such as drones, intelligent mobile spraying robots, etc.) in order to complete the spraying task efficiently and accurately in agricultural operations. This process is mainly based on the characteristics of the operation area, the performance of the spraying equipment and the operation requirements, and the best driving path is determined through algorithms and technical means; there is a close relationship between spraying path planning and agricultural technology extension services. As an important technology for intelligent agricultural operations, spraying path planning can significantly improve the utilization efficiency of pesticides with its high efficiency and accuracy; agricultural technology extension services are an important means to popularize advanced agricultural technologies to farmers and help them improve agricultural production efficiency; through agricultural technology extension services, farmers can understand the advantages and application methods of spraying path planning technology, so as to apply it in actual production and improve agricultural production efficiency and quality; at the same time, agricultural technology extension services can also continuously optimize and improve spraying path planning technology according to farmers' actual needs and feedback, so as to make it more adaptable to the actual needs of agricultural production; therefore, spraying path planning belongs to the concrete content of agricultural technology extension services.

[0003] The existing application is 201910169359.4, and the document named "Orchard Intelligent Self-propelled Spraying System and Control Method thereof" points out that the system includes a sprayer, a positioning base station, a handheld monitoring terminal and a driving controller. The sprayer is used to spray fruit trees with pesticides or foliar fertilizers. The positioning base station is set at the center of the orchard, and the driving controller is set on the sprayer; the driving controller is used to control the automatic driving and spraying operations of the sprayer. The control method of the present invention includes: a). Surveying and mapping operation points; b). Path planning; c). Automatic driving and spraying; d). Automatic dosing operation; The spray system and control method of the present invention realize that the sprayer drives according to the planned path to perform spraying or foliar fertilizer spraying operations, avoiding the harm of spraying agriculture to personnel health, avoiding heatstroke, improving spraying efficiency, and reducing labor costs; but it cannot control the amount of medicine used and ensure that the fruit trees or plants in the orchard are fully sprayed.

[0004] In combination with the above documents and existing technologies, the environment in existing orchards is complex, and there may be a large number of obstacles such as branches of fruit trees. Some obstacles will affect the moving path of the spraying equipment. Traditional spraying equipment is often unable to flexibly avoid obstacles, resulting in poor spraying effect or equipment damage; traditional spray pressure equipment often relies on manual experience or fixed settings to control the amount of spraying when performing spraying operations, and it is difficult to flexibly adjust according to the actual situation of the orchard; if excessive spraying or the spraying range is too large, it will not only waste medicine, but may also cause pollution to the orchard environment and surrounding ecosystems; if the spraying amount is small or the spraying range is too small, the coverage rate of the liquid medicine on the fruit trees will be reduced, and it will not have a good insecticidal effect. At this time, multiple spraying treatments are required, which greatly affects the spraying efficiency. Summary of the invention

[0005] 1. Technical issues to be resolved

[0006] In view of the deficiencies in the prior art, the present invention provides an intelligent spraying path planning system and method. The system integrates information collection, path planning, real-time identification and dynamic adjustment functions, realizes the comprehensive intelligence of orchard spraying operations, improves the accuracy and efficiency of spraying operations, and ensures the high efficiency and accuracy of spraying operations by automatically adjusting the estimated spray flow rate and the number of nozzles opened for each nozzle, while also ensuring that the fruit trees are fully covered with drugs. By accurately controlling the spraying amount and spray flow rate, the excessive use and dispersion of drugs are reduced, the pollution to the environment is reduced, and the problems raised in the background technology are solved.

[0007] (II) Technical solution

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0009] An intelligent spraying path planning system, comprising:

[0010] The information collection module, under the condition of obtaining the two-dimensional map of the target orchard, marks the dividing line on the side of any fruit tree planting area in the target orchard; uses image processing technology and edge detection algorithm to obtain the actual pest area; and maps the actual pest area onto the dividing line to generate a standard pest area;

[0011] The data analysis and processing module collects the area and number of vegetation in the actual pest area, simultaneously obtains the area of ​​the standard pest area, and obtains the vegetation density and pest severity value in the actual pest area through analysis and calculation; it is characterized by also including:

[0012] The path planning module extracts the coordinate set formed by the coordinates of each corner of the standard pest area under the condition that the target orchard entrance is determined as the origin coordinate, and uses the path planning algorithm to generate a traversal planning path; in the process of executing the traversal planning path, the spraying end uses the probe part on the spraying end to perform a real-time recognition mechanism, and issues a dynamic fine-tuning instruction based on the recognition result. The spraying end receives the fine-tuning instruction and changes the traversal planning path in real time;

[0013] The spraying adjustment module, when the spraying end enters the standard pest area, the nozzle assembly on the spraying end facing the standard pest area is pre-opened, and the nozzle opening number M is output based on the maximum height of the fruit tree obtained in real time by the spraying end and the pre-built rule engine, and the corresponding nozzle assembly executes the output result; extract and construct a guidance calculation model based on the nozzle opening number M, vegetation density and pest degree value, and generate the estimated spraying flow rate of a single nozzle;

[0014] The evaluation and adjustment module triggers the sampling survey mechanism after executing a traversal planning path, compares the average mortality rate obtained with the preset standard threshold, and decides whether to implement the correction and adjustment strategy based on the comparison results.

[0015] Further, the separation line is the central axis of a predetermined moving path in the target orchard; the image processing technology includes an image processing algorithm and an image segmentation algorithm;

[0016] The process of obtaining the actual infestation area is as follows:

[0017] Image preprocessing: Preprocess the two-dimensional image of the target orchard, including grayscale and denoising;

[0018] Pest feature extraction: Use image processing algorithms to extract features of pest areas;

[0019] Region segmentation: Based on the extracted features, the image segmentation algorithm is used to segment the pest area from the two-dimensional image;

[0020] Boundary determination: Perform boundary detection on the segmented pest area and use edge detection algorithm to obtain the boundary of the pest area, that is, the actual pest area.

[0021] Furthermore, image registration technology is used to map the actual pest area onto the separation line.

[0022] Furthermore, the analysis and calculation process of the vegetation density in the actual pest area is: divide the number of vegetation in the actual pest area by the area of ​​the actual pest area to obtain the vegetation density in the actual pest area; the analysis and calculation process of the pest severity value in the actual pest area is: divide the area of ​​the actual pest area by the area of ​​the standard pest area to obtain the pest severity value in the actual pest area.

[0023] Furthermore, the process of generating the traversal planning path is as follows:

[0024] Coordinate information digitization: converting each coordinate in the coordinate set into digital form;

[0025] Construct map markings: Based on the two-dimensional map of the target orchard, mark the standard pest area, fruit tree planting area and the entrance of the target orchard on the two-dimensional map;

[0026] Path planning algorithm: TSP algorithm is used to solve the shortest path.

[0027] Furthermore, the spraying robot walks according to the traversal planning path, and the matching probe parts on the spraying end include probe A and probe B, and probe A and probe B are symmetrically installed at the edges of both sides of the front end of the spraying end;

[0028] The process of implementing the real-time recognition mechanism is as follows:

[0029] The probe monitors in real time whether there are obstacles in the direction of the spraying end;

[0030] If one of the probe groups detects the presence of an obstacle, the recognition result is: the corresponding group of probes detects the obstacle;

[0031] If both sets of probes detect the existence of obstacles, the recognition result is: both sets of probes detect obstacles;

[0032] If no probe detects the existence of an obstacle, the recognition result is: no obstacle is detected;

[0033] When issuing a fine-tuning command based on the recognition result, if the recognition result is: the corresponding group of probes detects an obstacle;

[0034] The fine-tuning instruction is: the spray end is deflected toward the side away from the detected obstacle until both sets of probes detect no obstacles; if both sets of probes detect obstacles during the deflection process, the comparison strategy is run: identify the maximum diameter of the obstacles detected by probe A and probe B, and compare the maximum diameters of the two sets of obstacles; if probe A exceeds the maximum diameter detected by probe B, the spray end is deflected toward the side of probe B until probe A detects no obstacles; otherwise, the spray end is deflected toward the side of probe A until probe B detects no obstacles;

[0035] If the identification result is: both sets of probes detect fruit tree branches;

[0036] Then the content of the fine-tuning instruction is: run the comparison strategy;

[0037] If the identification result is: no fruit tree branches or trunks are detected;

[0038] The content of the fine-tuning instruction is: no response action.

[0039] Furthermore, there are two groups of nozzle assemblies, both of which are arranged at the head of the spraying end, and the two groups of nozzle assemblies are symmetrically distributed, both of which are quarter-circular structures, and a number of nozzles are evenly arranged on the arc surface of the nozzle assembly.

[0040] Furthermore, when the rule engine is run to output the number of nozzle openings M, the formula used is as follows:

[0041]

[0042] In the formula, H represents the maximum height of the fruit tree, H max , H min Respectively represent the highest and lowest heights of the corresponding type of fruit trees, M max Indicates the total number of sprinklers in the sprinkler assembly; Indicates rounding down;

[0043] The process of building a guidance calculation model to generate an estimated spray flow rate for a single sprinkler is as follows:

[0044] The extracted vegetation density and pest severity values ​​are normalized to the interval [0, 1];

[0045] Then, based on the normalized data and the number of nozzle openings M, the following formula is established:

[0046]

[0047] Where Q represents the estimated spray flow rate of a single nozzle, D represents the vegetation density, Ir represents the pest severity value, k1 and k2 are adjustment factors, and their value ranges are [0, 1].

[0048] Furthermore, the process of triggering the sampling survey mechanism is: randomly selecting a number of sample points in any pest-infested area, counting the pest mortality rate of the pests in each sample point, and calculating the average mortality rate; after comparing the obtained average mortality rate with the preset standard threshold;

[0049] If the comparison result is: the average mortality rate exceeds the standard threshold, no response action will be taken;

[0050] If the comparison result is: the average mortality rate does not exceed the standard threshold, the correction adjustment strategy is implemented;

[0051] The revised adjustment strategy is to construct a revised model based on the average mortality rate, the standard threshold, and the difference between the two to calculate the revised value. The process of constructing the revised model is as follows:

[0052]

[0053] In the formula, R represents the correction value, G represents the preset adjustment value, and Bz represents the standard threshold value. represents the average mortality rate, Δ represents the difference between the average mortality rate and the standard threshold;

[0054] The obtained correction value is added to the guidance calculation model constructed when executing a traversal planning path, so that the correction value changes the original single nozzle estimated spray flow rate by accumulation; the operations in the evaluation and adjustment module are executed cyclically until the comparison result is: the average mortality rate exceeds the standard threshold.

[0055] An intelligent spraying path planning method includes the following steps:

[0056] S1. Under the condition of obtaining a two-dimensional map of the target orchard, mark a dividing line on the side of any fruit tree planting area in the target orchard; use image processing technology and edge detection algorithm to obtain the actual pest area; and map the actual pest area onto the dividing line to generate a standard pest area;

[0057] S2. Collect the area and number of vegetation in the actual pest area, and simultaneously obtain the area of ​​the standard pest area, and obtain the vegetation density and pest severity value in the actual pest area through analysis and calculation;

[0058] S3. Under the condition that the entrance of the target orchard is determined as the origin coordinate, the coordinate set formed by the coordinates of each corner of the standard pest area is extracted, and the path planning algorithm is used to generate a traversal planning path; in the process of executing the traversal planning path, the spraying end uses the matching probe on the spraying end to execute a real-time recognition mechanism, and issues a dynamic fine-tuning instruction based on the recognition result. The spraying end receives the fine-tuning instruction and changes the traversal planning path in real time;

[0059] S4. When the spraying end enters the standard pest area, the spray head assembly on the spraying end facing the standard pest area is pre-opened, and the number of nozzle openings M is outputted synchronously based on the maximum height of the fruit tree obtained in real time by the spraying end and the pre-built rule engine, and the corresponding nozzle assembly executes the output result; extract and construct a guidance calculation model based on the number of nozzle openings M, vegetation density and pest severity value, and generate an estimated spray flow rate of a single nozzle;

[0060] S5. After executing a traversal planning path, the sampling survey mechanism is triggered, and the average mortality rate obtained is compared with the preset standard threshold. According to the comparison result, it is decided whether to implement the correction adjustment strategy.

[0061] (III) Beneficial effects

[0062] The present invention provides an intelligent spraying path planning system and method, which has the following beneficial effects:

[0063] (1) This solution uses image processing technology and edge detection algorithms to accurately identify the pest areas in the orchard and map them to the preset dividing lines to generate standard pest areas, thereby improving the accuracy of pest identification and positioning precision and providing a reliable data basis for subsequent processing. By collecting the area of ​​the actual pest area, the number of vegetation, and the area of ​​the standard pest area, the system can analyze and calculate the vegetation density and pest severity value in the actual pest area, which provides orchard managers with quantitative pest assessment indicators and helps to formulate more scientific and reasonable pest control strategies.

[0064] (2) This solution uses a path planning algorithm, combined with a set of corner coordinates of a standard pest area, to quickly generate the shortest path from the orchard entrance through all pest areas and back to the entrance. In the process of executing the traversal planning path, obstacles such as branches of fruit trees in the forward direction are detected through a real-time recognition mechanism, and dynamic fine-tuning instructions are issued based on the recognition results. This enables the spraying robot to flexibly respond to the complex environment in the orchard, avoid collisions with obstacles to the greatest extent possible, and ensure that the pest area is fully sprayed, preliminarily guaranteeing the coverage rate. The system integrates information collection, path planning, real-time recognition, and dynamic adjustment functions, realizing the comprehensive intelligence of orchard spraying operations, which not only improves the accuracy and efficiency of spraying operations, but also reduces labor costs and labor intensity.

[0065] (3) This solution dynamically adjusts the number of nozzles opened according to the maximum height of the fruit trees through the rule engine, realizing intelligent adjustment of the spraying amount, which not only ensures that the fruit trees are fully covered by the medicine, but also avoids unnecessary waste of medicine. By constructing a guidance calculation model, combined with the vegetation density and pest severity value, the estimated spraying flow rate of each nozzle is accurately calculated, ensuring the efficiency and accuracy of the spraying operation, and to a certain extent, maximizing the efficiency of drug use. By accurately controlling the spraying amount and spraying flow rate, the excessive use and drift of drugs are reduced, reducing pollution to the environment.

[0066] (4) This scheme dynamically adjusts the spraying strategy based on the comparison results between the average mortality rate and the preset standard threshold, ensuring that the spraying operation can be flexibly optimized according to the actual effect, thereby improving the efficiency and accuracy of the spraying. In conjunction with the constructed correction model, the correction value is calculated and added to the guidance calculation model, thereby achieving precise control of the estimated spray flow rate of a single nozzle, further avoiding drug waste and excessive use, and ensuring that the fruit trees are fully covered by the drug. The operations in the evaluation and adjustment module are executed cyclically until the average mortality rate exceeds the standard threshold. This continuous optimization mechanism ensures the continuous improvement of the spraying effect and meets the actual needs of orchard management. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1It is a schematic diagram of the system module of the present invention;

[0068] Figure 2 It is a schematic diagram of the overall process of the present invention;

[0069] Figure 3 is a local two-dimensional scene graph of the target orchard in the present invention;

[0070] Figure 4 It is a schematic diagram of the application scenario of the probe component in the present invention;

[0071] Figure 5 It is a schematic diagram of the application scenario of the nozzle assembly of the present invention;

[0072] Figure numerals: 1. Nozzle assembly. DETAILED DESCRIPTION

[0073] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0074] Embodiment 1:

[0075] See also Figure 1 , Figures 3 to 5 This embodiment provides an intelligent spraying path planning system. The application scenarios of this system are mostly used in greenhouses for planting fruit trees or fruits and vegetables. It can also be used in areas where fruit trees are cultivated according to needs. Modern fruit tree planting and cultivation are mostly arranged according to established areas, which is convenient for subsequent observation, spraying and picking;

[0076] The system consists of several functional modules, each of which operates in sequence, namely, information collection module, data analysis and processing module, path planning module, spraying adjustment module, and evaluation and adjustment module;

[0077] Now each of the above functional modules is described and elaborated in turn:

[0078] Information collection module:

[0079] Under the condition of obtaining the two-dimensional map of the target orchard, a dividing line is marked on the side of any fruit tree planting area in the target orchard; the actual pest area is obtained by using image processing technology and edge detection algorithm; and the actual pest area is mapped onto the dividing line to generate a standard pest area; wherein the dividing line is the central axis of the predetermined moving path in the target orchard;

[0080] The two-dimensional map of the target orchard is a plan view, which can be obtained by drone aerial photography and satellite image capture;

[0081] Drone aerial photography:

[0082] Use drones with high-resolution cameras to take aerial photos of orchards and obtain bird's-eye views of the orchards. This method is flexible and efficient and can quickly cover the entire orchard area. Import the orchard images taken by drones into professional image processing software (such as DJI Maps) and generate a two-dimensional plane of the orchard through image stitching, correction and other processing steps. This process may involve image sorting and confirmation to ensure that the images are clear, not exposed, and not missing.

[0083] Satellite image capture:

[0084] Utilize existing satellite image resources, such as Google Earth and Baidu Maps, which provide rich geographic information, including satellite images of orchard areas. Locate the target orchard area on the satellite image platform, use the image capture tool to capture the image of the orchard, and perform appropriate cropping and adjustments to obtain a clear two-dimensional image of the orchard.

[0085] In this embodiment, drone aerial photography can be used to obtain and store a two-dimensional image of the target orchard in advance;

[0086] Reference Figure 3 As shown, the fruit trees in the target orchard need to be planned in advance before planting. For the selection of fruit tree planting areas, they are all arranged in several vertical patterns, and the areas between adjacent fruit tree planting areas are for subsequent spraying or staff to enter. The distance between adjacent fruit tree planting areas is consistent for easy management.

[0087] As for the dividing line, the dividing line is a virtual route, which can be displayed in the two-dimensional map of the target orchard as part of the traversal planning path of the spraying end; the central axis of the established moving path in the target orchard indicates that the path formed between adjacent fruit tree planting areas has a certain width, and the center line of the path is the central axis that divides the two symmetrical path areas; for the dividing line at the edge of the fruit tree planting area, although it is not adjacent to another fruit tree planting area, its path specifications are consistent, so the various dividing lines in the same target orchard are always evenly distributed;

[0088] Image processing technology includes image processing algorithms and image segmentation algorithms;

[0089] The process of obtaining the actual infestation area is as follows:

[0090] Image preprocessing: Preprocess the two-dimensional image of the target orchard, including grayscale, denoising, and contrast enhancement, to improve image quality and facilitate subsequent processing; Pest feature extraction: Use image processing algorithms (including color space conversion and texture analysis) to extract the features of the pest area; for example, the pest area presents a different color (including yellow and brown) or texture (including leaf damage and spots) from normal vegetation; Region segmentation: Based on the extracted features, use image segmentation algorithms (including threshold segmentation, region growing, and cluster segmentation) to segment the pest area from the two-dimensional image, which may require multiple iterations and parameter adjustments to obtain accurate segmentation results; Boundary determination: Perform boundary detection on the segmented pest area, and use edge detection algorithms (including Canny and Sobel) to obtain the accurate boundary of the pest area, that is, the actual pest area;

[0091] Image registration techniques are used when mapping the actual infestation area onto the divider lines;

[0092] Align the coordinates of the actual infested area with the standard coordinates on the dividing line by calculating the coordinates of the centroid or bounding box of the infested area and converting them to the corresponding coordinates on the dividing line; refer to Figure 3 It can be seen that the range of the actual pest area and its corresponding standard pest area in the actual detection;

[0093] The shaded area is the actual pest infestation area;

[0094] The rectangular part with four corner coordinates formed by the dividing line is the standard pest area.

[0095] Data processing and analysis module:

[0096] Collect the area and number of vegetation in the actual pest area, and simultaneously obtain the area of ​​the standard pest area. After analysis and calculation, the vegetation density and pest severity value in the actual pest area are obtained.

[0097] The area of ​​the actual pest area is accurately measured using a measuring tool (such as GPS, tape measure, etc.) or a geographic information system (GIS) software. In this embodiment, in order to ensure automated measurement, the geographic information system software is used for measurement and collection. The number of vegetation is the number of vegetation in the actual pest area, which can be directly obtained from the two-dimensional map of the target orchard. The area of ​​the standard pest area can be obtained based on the area of ​​the actual pest area in combination with the geographic information system software.

[0098] The analysis and calculation process of the vegetation density in the actual pest area is as follows: divide the number of vegetation in the actual pest area by the area of ​​the actual pest area, and the result is the vegetation density in the actual pest area; the pest severity value in the actual pest area is used to quantify the impact of pests on fruit trees, which is expressed by the proportion of the pest area. The analysis and calculation process of the pest severity value in the actual pest area is as follows: divide the area of ​​the actual pest area by the area of ​​the standard pest area, and the result is the pest severity value in the actual pest area.

[0099] Specifically, by using image processing technology and edge detection algorithms, it is possible to accurately identify pest areas in the orchard and map them to preset dividing lines to generate standard pest areas, which improves the accuracy of pest identification and positioning precision, and provides a reliable data basis for subsequent processing; by collecting the area of ​​the actual pest area, the number of vegetation, and the area of ​​the standard pest area, the system can analyze and calculate the vegetation density and pest severity in the actual pest area, which provides orchard managers with quantitative pest assessment indicators and helps to formulate more scientific and reasonable pest and disease control strategies.

[0100] Path planning module:

[0101] Under the condition that the entrance of the target orchard is determined as the origin coordinate, the coordinate set formed by the coordinates of each corner of the standard pest area is extracted, and the path planning algorithm is used to generate the traversal planning path; in the process of executing the traversal planning path, the spraying end uses the matching probe on the spraying end to execute the real-time recognition mechanism, and issues dynamic fine-tuning instructions based on the recognition results. The spraying end receives the fine-tuning instructions and changes the traversal planning path in real time;

[0102] Among them, refer to Figure 3 As shown, the standard pest area is a rectangle, and the four right-angled sides of the rectangle are the corners. Therefore, there are four sets of corner coordinates for one standard pest area, and eight sets of corner coordinates for two standard pest areas. Figure 3 These eight groups of edges and corners together form a coordinate set;

[0103] The path planning algorithm adopts the traveling salesman problem (TSP) algorithm or its variants, such as genetic algorithm and ant colony algorithm;

[0104] The process of generating a traversal planning path is as follows:

[0105] Coordinate information digitization: The coordinates in the image coordinate set (in Figure 3For example, the coordinates in the coordinate set (including coordinates 1 to 8) are converted into digital form for computer processing, and the entrance of the target orchard is determined as the starting point (i.e., the origin coordinates); map marking is constructed: based on the two-dimensional map of the target orchard, the standard pest area, the fruit tree planting area and the entrance of the target orchard are marked on the two-dimensional map; path planning algorithm: the traveling salesman problem (TSP) algorithm or its variants, such as genetic algorithm, ant colony algorithm, etc., are used to solve the shortest path problem, with the input being the origin coordinates and the coordinate set, and the output being the shortest path starting from the entrance of the target orchard, traversing each coordinate in the coordinate set, and returning to the entrance of the target orchard;

[0106] For example:

[0107] Coordinate 1 can be expressed as (x1, y1), coordinate 2 as (x2, y2), and so on. The coordinates of the target orchard entrance can be set to (0, 0) or determined according to actual needs. On the two-dimensional plane, use different colors or symbols to mark the standard pest area, fruit tree planting area and target orchard entrance to ensure that the map is consistent with the actual orchard layout. Select a suitable algorithm for solving. For example, the genetic algorithm searches for the optimal solution by simulating natural selection and genetic mechanisms. The ant colony algorithm finds the shortest path by simulating ant foraging behavior. The specific implementation of the algorithm needs to be adjusted according to the size of the orchard and the number of coordinate points.

[0108] Reference Figure 3 As shown, the spraying end adopts a spraying robot that can move autonomously;

[0109] The spraying robot walks according to the traversal planning path. The matching probe parts on the spraying end include two sets of probes A and B. Probes A and probes B are respectively installed at the positions of the two side edges of the front end of the spraying end (not marked in the figure). The probe parts are used to detect whether there are extra branches of fruit trees (i.e. obstacles) in the forward direction of the spraying end and the maximum diameter of the branches of fruit trees (i.e. the maximum width);

[0110] The process of implementing the real-time recognition mechanism is as follows:

[0111] The probe component monitors in real time whether there are any extra branches of fruit trees in the direction of the spraying end;

[0112] If one of the probe groups detects the existence of a fruit tree branch, the identification result is: the corresponding group of probes detects the existence of a fruit tree branch;

[0113] If both sets of probes detect the existence of fruit tree branches, the identification result is: both sets of probes detect the existence of fruit tree branches;

[0114] If no probe detects the existence of a fruit tree branch, the identification result is: no fruit tree branch is detected;

[0115] When dynamic fine-tuning instructions are issued based on the recognition result, if the recognition result is: the corresponding group of probes detects the branches of fruit trees;

[0116] The content of the fine-tuning instruction is: the spraying end is deflected toward the side away from the monitored fruit tree branches until both sets of probes fail to detect the fruit tree branches; if both sets of probes detect the fruit tree branches during the deflection process, the comparison strategy is run: identify the maximum branch width of the fruit tree branches monitored by probes A and B, and compare the maximum branch widths of the two sets; if probe A exceeds the maximum branch width of probe B, the spraying end is deflected toward the side of probe B until probe A fails to detect the fruit tree branches; if probe B exceeds the maximum branch width of probe A, the spraying end is deflected toward the side of probe A until probe B fails to detect the fruit tree branches;

[0117] It should be noted that in actual use, the probes configured on both sides of the front end of the spraying end monitor the two sides of the front edge of the spraying end respectively. Figure 4 It can be seen that the specific positions of probe A and probe B, probe A and B have the same structure, function and principle, so taking probe A as an example, the ultrasonic probe part in probe A is realized by pre-adjusting its transmission angle and receiving sensitivity, referring to Figure 4 As shown in the figure, the detection range is 5 meters in front and the maximum width of the object is 2 meters. Then, when necessary, the camera probe part of the probe A is used to detect and analyze the maximum width of the branches of the fruit tree. Therefore, the probe A can realize the function of detecting obstacles and judging the size of obstacles.

[0118] The technical principle of the maximum branch width detected by any probe in the fruit tree branches is:

[0119] The camera probe part analyzes the maximum diameter of obstacles among the obstacles in front, mainly through image processing technology; first, the camera probe part captures the image in front and identifies the outline of the obstacle through the edge detection algorithm; then, the feature extraction technology is used to obtain the key feature points of the obstacle, such as corner points, intersections, etc.; then, by calculating the distance between these feature points, the size of the obstacle, including its maximum diameter, can be estimated.

[0120] If the identification result is: both sets of probes detect fruit tree branches;

[0121] Then the content of the fine-tuning instruction is: run the comparison strategy;

[0122] If the identification result is: no fruit tree branches or trunks are detected;

[0123] The content of the fine-tuning instruction is: no response action.

[0124] Specifically, the path planning algorithm, combined with the corner coordinate set of the standard pest area, can quickly generate the shortest path from the orchard entrance to traverse all pest areas and return to the entrance. In the process of executing the traversal planning path, the real-time recognition mechanism detects obstacles such as fruit tree branches in the forward direction, and issues dynamic fine-tuning instructions based on the recognition results. This enables the spraying robot to flexibly respond to the complex environment in the orchard, avoid collisions with obstacles to the greatest extent, and ensure that the pest area is fully sprayed, initially guaranteeing the coverage rate.

[0125] The probe equipped on the spraying robot can monitor obstacles in front in real time and accurately analyze the size of the obstacles through image processing technology. When an obstacle is detected, the robot can perform precise avoidance operations based on the position and size of the obstacle, improving the safety and efficiency of the spraying operation. The system integrates multiple functional modules such as information collection, path planning, real-time recognition, and dynamic adjustment, realizing the comprehensive intelligence of orchard spraying operations, which not only improves the accuracy and efficiency of spraying operations, but also reduces labor costs and labor intensity.

[0126] Spraying adjustment module:

[0127] Under the condition that the spraying end enters the standard pest area, the spray head assembly 1 on the spraying end facing the standard pest area is pre-opened (indicating the state of being ready to be opened but not yet opened), and the number of nozzle openings M is outputted synchronously based on the maximum height of the fruit tree obtained in real time by the spraying end and the pre-built rule engine, and the corresponding spray head assembly 1 executes the output result; extract and construct a guidance calculation model based on the number of nozzle openings M, vegetation density and pest degree value, and generate an estimated spray flow rate of a single nozzle (i.e., the spray volume per minute, unit: liter / minute);

[0128] Among them, refer to Figure 5 It can be seen that there are two groups of nozzle assemblies 1, both of which are arranged at the head (i.e., the front end) of the spraying robot (i.e., the spraying end), and the two groups of nozzle assemblies 1 are symmetrically distributed, both of which are quarter-circular structures, and a number of nozzles are evenly arranged on the arc surface of the nozzle assembly 1 for spraying atomized insecticides; at the same time Figure 5 The approximate spraying range of each nozzle during operation is also given in ; M is a positive integer, Figure 5 As shown in , the maximum value of M is 5;

[0129] The maximum height of the fruit tree obtained in real time by the spraying end is obtained by using the ultrasonic sensor configured on the spraying end (refer to Figure 5 The position of the ultrasonic sensor can be clearly seen. The ultrasonic sensor is connected to the top of the fruit tree. The ultrasonic sensor transmits ultrasonic waves to the fruit tree and receives the reflected signal to generate the maximum height of the fruit tree.

[0130] When running the rule engine to output the number of nozzle openings M, the formula used is as follows:

[0131]

[0132] In the formula, H represents the maximum height of the fruit tree, H max , H min They represent the highest and lowest heights of the corresponding type of fruit trees. For example, if the highest height of type X fruit trees is 0 to 10 meters, then H max =10,H min =0,M max Indicates the total number of nozzles in nozzle assembly 1; Indicates rounding down;

[0133] Logical explanation: In the proportion calculation in the formula, when the lowest height is 0, the inverse of the ratio of the maximum height H of the fruit tree to the highest height is calculated. When the maximum height H of the fruit tree is close to 0, this ratio will become very large; when the maximum height H of the fruit tree is close to the highest height, this ratio is close to 1; the lowest height is used as a scaling factor to enlarge the ratio so as to distribute the value of M in a wider range; +1 is to ensure that the result is at least 1 to avoid working without a sprinkler head;

[0134] The process of building a guidance calculation model to generate an estimated spray flow rate for a single sprinkler is as follows:

[0135] The extracted vegetation density and pest severity values ​​are normalized to the interval [0, 1];

[0136] Then, based on the normalized data and the number of nozzle openings M, the following formula is established:

[0137]

[0138] In the formula, Q represents the estimated spray flow rate of a single nozzle, D represents the vegetation density, Ir represents the pest severity value, k1 and k2 are both adjustment factors used to adjust the influence of vegetation density and pest severity on the spray flow rate, and their value ranges are [0, 1];

[0139] Logic explanation: The nozzle opening number M is taken as part of the basic flow, directly multiplied by the nonlinear function result behind, and then calculated by ln(1+D*e k1*Ir ) is introduced in the form of . When the vegetation density increases, the spray flow rate will also increase, but the rate of increase will gradually slow down with the increase of density, so as to avoid the waste of a large amount of liquid medicine due to the inability to enter the deep area. Combined with the vegetation density, the exponential function e k1*IrTo magnify or reduce the effect of vegetation density on flow, the higher the pest level, the faster the spray flow increases; k1 adjusts the sensitivity of pest level to the effect of vegetation density on flow; k2 is used to adjust the degree of reduction in spray flow when vegetation density is low or pest level is low; 1+k2*(1-D)*(1-Ir) is used to ensure that when vegetation density or pest level is very low, spray flow will not increase indefinitely, but has an upper limit;

[0140] Overall description:

[0141] The number of nozzles opened is M = 3;

[0142] Vegetation density D = 0.6 (assumed to be normalized);

[0143] The pest severity value Ir = 0.8 (assuming normalization);

[0144] Adjustment factors k1=0.5, k2=0.3;

[0145] Substitute into the formula to calculate: Q = (3*ln(1+0.6*1.4818)) / (1+0.024);

[0146] Q = (3*ln(1.8951) / 1.024)≈1.89 (keep two decimal places) (unit: liters / minute).

[0147] Specifically, the system dynamically adjusts the number of nozzles opened according to the maximum height of the fruit trees through the rule engine, realizing intelligent adjustment of the spraying amount, which not only ensures that the fruit trees are fully covered by the medicine, but also avoids unnecessary waste of medicine. By building a guiding calculation model, combined with the vegetation density and pest severity value, the estimated spraying flow rate of each nozzle is accurately calculated, which ensures the efficiency and accuracy of the spraying operation, and to a certain extent, maximizes the efficiency of the use of medicine.

[0148] This technical solution can be flexibly adjusted according to different orchard environments (such as tree height, vegetation density, degree of insect pests, etc.), and has strong adaptability and versatility; by precisely controlling the spraying amount and spraying flow, it reduces excessive use and dispersion of drugs, and to a certain extent minimizes the use of drugs, thereby reducing pollution to the environment.

[0149] Assessment and Adjustment Module:

[0150] After executing a traversal planning path, the sampling survey mechanism is triggered to compare the average mortality rate obtained with the preset standard threshold, and decide whether to implement the correction adjustment strategy based on the comparison result;

[0151] The process of triggering the sampling survey mechanism is as follows:

[0152] Randomly select a number of sample points in any pest area, count the pest survival status in each sample point (i.e. the pest mortality rate corresponding to each sample point), and calculate the average mortality rate;

[0153] The specific numerical calculation process is as follows:

[0154] The pest mortality rate corresponding to each sampling point = (the number of pests that have fallen off + the number of pests that died but did not fall off) / the total number of pests × 100%; for example, if an inspection is conducted 24 hours after spraying a pest-infested area, it is found that there are 100 pests that have fallen off and 50 pests that died but did not fall off, and the total number of pests in the area is 200, then the pest mortality rate is:

[0155] Pest mortality rate = (100 + 50) / 200 × 100% = 75%;

[0156] Then the pest mortality rate of each sample point is averaged and if the calculated result is still 75%, it means that the average mortality rate after spraying has reached 75%, which can be used as a specific value to evaluate the effect of spraying; the higher the mortality rate, the better the effect of spraying.

[0157] If the comparison result is: the average mortality rate exceeds the standard threshold, no response action will be taken;

[0158] If the comparison result is: the average mortality rate does not exceed the standard threshold, the correction adjustment strategy is implemented;

[0159] The revised adjustment strategy is as follows:

[0160] Based on the average mortality rate, the standard threshold and the difference between the two, a correction model is constructed to calculate the correction value. The process of constructing the correction model is as follows:

[0161]

[0162] In the formula, R represents the correction value, G represents the preset adjustment value, usually with an increment of 0.5 liters / minute as a unit, and G is also used as a correction coefficient to control the amplitude of the correction value. Bz represents the standard threshold value. represents the average mortality rate;

[0163] Logical explanation: First, calculate the difference Δ between the average mortality rate and the standard threshold. This difference directly reflects the gap between the current mortality rate and the expected mortality rate (i.e., the standard threshold). Introduce the correction coefficient G, which is a small positive number used to control the overall amplitude of the correction value. The selection of K should be based on practical applications, usually a unit flow increment, to ensure that the corrected spray flow rate can effectively adjust the mortality rate without causing waste or excess. Directly take the smaller value between the difference and 1 to calculate the correction value to ensure that the correction value is not too large.

[0164] The obtained correction value is added to the guidance calculation model constructed when executing a traversal planning path, so that the correction value changes the original single nozzle estimated spray flow rate by accumulation. The specific formula is as follows:

[0165]

[0166] The operations in the evaluation and adjustment module are executed in a loop until the comparison result is that the average mortality rate exceeds the standard threshold.

[0167] Specifically, based on the comparison results between the average mortality rate obtained by the sampling survey mechanism and the preset standard threshold, the spraying strategy is dynamically adjusted, which ensures that the spraying operation can be flexibly optimized according to the actual effect, thereby improving the efficiency and accuracy of spraying; in conjunction with the constructed correction model, the correction value is calculated and added to the guidance calculation model, thereby achieving precise control of the estimated spray flow rate of a single nozzle, further achieving the effect of avoiding drug waste and excessive use, while ensuring that the fruit trees are fully covered by the drug, and cyclically executing the operations in the evaluation and adjustment module until the average mortality rate exceeds the standard threshold. This continuous optimization mechanism ensures the continuous improvement of the spraying effect and meets the actual needs of orchard management.

[0168] Embodiment 2:

[0169] See also Figure 2 Based on Example 1, this embodiment also provides an intelligent spraying path planning method, including the following specific steps:

[0170] S1. Under the condition of obtaining a two-dimensional map of the target orchard, mark a dividing line on the side of any fruit tree planting area in the target orchard; use image processing technology and edge detection algorithm to obtain the actual pest area; and map the actual pest area onto the dividing line to generate a standard pest area;

[0171] S2. Collect the area and number of vegetation in the actual pest area, and simultaneously obtain the area of ​​the standard pest area, and obtain the vegetation density and pest severity value in the actual pest area through analysis and calculation;

[0172] S3. Under the condition that the entrance of the target orchard is determined as the origin coordinate, the coordinate set formed by the coordinates of each corner of the standard pest area is extracted, and the path planning algorithm is used to generate a traversal planning path; in the process of executing the traversal planning path, the spraying end uses the matching probe on the spraying end to execute a real-time recognition mechanism, and issues a dynamic fine-tuning instruction based on the recognition result. The spraying end receives the fine-tuning instruction and changes the traversal planning path in real time;

[0173] S4. When the spraying end enters the standard pest area, the spray head assembly 1 on the spraying end facing the standard pest area is pre-opened, and the number of nozzle openings M is outputted synchronously based on the maximum height of the fruit tree obtained in real time by the spraying end and the pre-built rule engine, and the corresponding spray head assembly 1 executes the output result; extract and construct a guidance calculation model based on the number of nozzle openings M, vegetation density and pest degree value, and generate an estimated spray flow rate of a single nozzle;

[0174] S5. After executing a traversal planning path, the sampling survey mechanism is triggered, and the average mortality rate obtained is compared with the preset standard threshold. According to the comparison result, it is decided whether to implement the correction adjustment strategy.

[0175] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product. A person of ordinary skill in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution.

[0176] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0177] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. An intelligent spraying path planning system, the system comprising: The information collection module, under the condition of obtaining a two-dimensional map of the target orchard, marks a dividing line on the side of any fruit tree planting area in the target orchard; Using image processing technology and edge detection algorithm, the actual pest area is obtained; and the actual pest area is mapped onto the dividing line to generate a standard pest area; The data analysis and processing module collects the area and number of vegetation in the actual pest area, simultaneously obtains the area of ​​the standard pest area, and obtains the vegetation density and pest severity value in the actual pest area through analysis and calculation; it is characterized by also including: The path planning module extracts the coordinate set formed by the coordinates of each corner of the standard pest area under the condition that the target orchard entrance is determined as the origin coordinate, and uses the path planning algorithm to generate a traversal planning path; in the process of executing the traversal planning path, the spraying end uses the probe part on the spraying end to perform a real-time recognition mechanism, and issues a dynamic fine-tuning instruction based on the recognition result. The spraying end receives the fine-tuning instruction and changes the traversal planning path in real time; The spraying adjustment module, when the spraying end enters the standard pest area, the spray head assembly (1) on the spraying end facing the standard pest area is pre-opened, and the spray head opening number M is outputted synchronously based on the maximum height of the fruit tree obtained in real time by the spraying end and the pre-built rule engine, and the corresponding spray head assembly (1) executes the output result; extracts and constructs a guidance calculation model based on the spray head opening number M, the vegetation density and the pest degree value, and generates an estimated spray flow rate of a single spray head; The evaluation and adjustment module triggers the sampling survey mechanism after executing a traversal planning path, compares the average mortality rate obtained with the preset standard threshold, and decides whether to implement the correction and adjustment strategy based on the comparison results.

2. The intelligent spraying path planning system according to claim 1, characterized in that: The separation line is the central axis of the predetermined moving path in the target orchard; the image processing technology includes image processing algorithm and image segmentation algorithm; The process of obtaining the actual infestation area is as follows: Image preprocessing: Preprocess the two-dimensional image of the target orchard, including grayscale and denoising; Pest feature extraction: Use image processing algorithms to extract features of pest areas; Region segmentation: Based on the extracted features, the image segmentation algorithm is used to segment the pest area from the two-dimensional image; Boundary determination: Perform boundary detection on the segmented pest area and use edge detection algorithm to obtain the boundary of the pest area, that is, the actual pest area.

3. The intelligent spraying path planning system according to claim 1, characterized in that: Image registration techniques are used to map the actual infestation area onto the divider lines.

4. The intelligent spraying path planning system according to claim 1, characterized in that: The analysis and calculation process of the vegetation density in the actual pest area is as follows: divide the number of vegetation in the actual pest area by the area of ​​the actual pest area to obtain the vegetation density in the actual pest area; the analysis and calculation process of the pest severity value in the actual pest area is as follows: divide the area of ​​the actual pest area by the area of ​​the standard pest area to obtain the pest severity value in the actual pest area.

5. The intelligent spraying path planning system according to claim 1, characterized in that: The process of generating a traversal planning path is as follows: Coordinate information digitization: converting each coordinate in the coordinate set into digital form; Construct map markings: Based on the two-dimensional map of the target orchard, mark the standard pest area, fruit tree planting area and the entrance of the target orchard on the two-dimensional map; Path planning algorithm: TSP algorithm is used to solve the shortest path.

6. The intelligent spraying path planning system according to claim 5, characterized in that: The spraying robot walks according to the traversal planning path. The matching probe parts on the spraying end include probe A and probe B, and probe A and probe B are symmetrically installed at the edges of both sides of the front end of the spraying end; The process of implementing the real-time recognition mechanism is as follows: The probe monitors in real time whether there are obstacles in the direction of the spraying end; If one of the probe groups detects the presence of an obstacle, the recognition result is: the corresponding group of probes detects the obstacle; If both sets of probes detect the existence of obstacles, the recognition result is: both sets of probes detect obstacles; If no probe detects the existence of an obstacle, the recognition result is: no obstacle is detected; When issuing a fine-tuning command based on the recognition result, if the recognition result is: the corresponding group of probes detects an obstacle; Then the fine-tuning instruction is: the spray end is deflected toward the side away from the detected obstacle until both sets of probes do not detect the obstacle; If both sets of probes detect obstacles during the deflection process, the comparison strategy will be run: identify the maximum diameter of the obstacles detected by probe A and probe B, and compare the maximum diameters of the two sets of obstacles; if probe A exceeds the maximum diameter detected by probe B, the spray end will deflect toward the side of probe B until probe A detects no obstacles; Otherwise, the spray end will deflect toward the side of probe A until probe B detects no obstacle; If the identification result is: both sets of probes detect fruit tree branches; Then the content of the fine-tuning instruction is: run the comparison strategy; If the identification result is: no fruit tree branches or trunks are detected; The content of the fine-tuning instruction is: no response action.

7. The intelligent spraying path planning system according to claim 6, characterized in that: The number of the nozzle assemblies (1) is two groups, both of which are arranged at the head of the spraying end, and the two groups of nozzle assemblies (1) are symmetrically distributed, both of which are quarter-circular structures, and a plurality of nozzles are evenly arranged on the arc surface of the nozzle assembly (1).

8. The intelligent spraying path planning system according to claim 1, characterized in that: When running the rule engine to output the number of nozzle openings M, the formula used is as follows: In the formula, H represents the maximum height of the fruit tree, H max , H min Respectively represent the highest and lowest heights of the corresponding type of fruit trees, M max Indicates the total number of nozzles in the nozzle assembly (1); Indicates rounding down; The process of building a guidance calculation model to generate an estimated spray flow rate for a single sprinkler is as follows: The extracted vegetation density and pest severity values ​​are normalized to the interval [0, 1]; Then, based on the normalized data and the number of nozzle openings M, the following formula is established: Where Q represents the estimated spray flow rate of a single nozzle, D represents the vegetation density, Ir represents the pest severity value, k1 and k2 are adjustment factors, and their value ranges are [0, 1].

9. The intelligent spraying path planning system according to claim 1, characterized in that: The process of triggering the sampling survey mechanism is: randomly select a number of sample points in any pest area, count the pest mortality rate of the pests in each sample point, and calculate the average mortality rate; compare the obtained average mortality rate with the preset standard threshold; If the comparison result is: the average mortality rate exceeds the standard threshold, no response action will be taken; If the comparison result is: the average mortality rate does not exceed the standard threshold, the correction adjustment strategy is implemented; The revised adjustment strategy is to construct a revised model based on the average mortality rate, the standard threshold, and the difference between the two to calculate the revised value. The process of constructing the revised model is as follows: In the formula, R represents the correction value, G represents the preset adjustment value, and Bz represents the standard threshold value. represents the average mortality rate, Δ represents the difference between the average mortality rate and the standard threshold; The obtained correction value is added to the guidance calculation model constructed when executing a traversal planning path, so that the correction value changes the original single nozzle estimated spray flow rate by accumulation; the operations in the evaluation and adjustment module are executed cyclically until the comparison result is: the average mortality rate exceeds the standard threshold.

10. An intelligent spraying path planning method, using any system described in claims 1 to 9, characterized in that: The steps include: S1. Under the condition of obtaining a two-dimensional map of the target orchard, mark a dividing line on the side of any fruit tree planting area in the target orchard; use image processing technology and edge detection algorithm to obtain the actual pest area; and map the actual pest area onto the dividing line to generate a standard pest area; S2. Collect the area and number of vegetation in the actual pest area, and simultaneously obtain the area of ​​the standard pest area, and obtain the vegetation density and pest severity value in the actual pest area through analysis and calculation; S3, under the condition that the entrance of the target orchard is determined as the origin coordinate, extract the coordinate set formed by the coordinates of each corner of the standard pest area, and use the path planning algorithm to generate a traversal planning path; When the spraying end is executing the traversal planning path, the probe part equipped on the spraying end is used to execute the real-time recognition mechanism, and a dynamic fine-tuning instruction is issued according to the recognition result. The spraying end receives the fine-tuning instruction and changes the traversal planning path in real time. S4. When the spraying end enters the standard pest area, the spray head assembly (1) on the spraying end facing the standard pest area is pre-opened, and the number of nozzle openings M is outputted synchronously based on the maximum height of the fruit tree obtained in real time by the spraying end and the pre-built rule engine, and the corresponding spray head assembly (1) executes the output result; extracts and constructs a guidance calculation model based on the number of nozzle openings M, vegetation density and pest severity value, and generates an estimated spray flow rate of a single nozzle; S5. After executing a traversal planning path, the sampling survey mechanism is triggered, and the average mortality rate obtained is compared with the preset standard threshold. According to the comparison result, it is decided whether to implement the correction adjustment strategy.

Images