Spraying parameter acquisition method based on plant canopy density and spraying machine

Abstract

The application provides a spray parameter acquisition method based on plant canopy density and a sprayer, and relates to the technical field of agricultural plant protection. The method comprises the following steps: determining the point cloud quantity of a plant to be sprayed in multiple canopy regions according to the canopy point cloud data of the plant to be sprayed; determining the leaf area index of each canopy region of the plant to be sprayed according to the point cloud quantity of the plant to be sprayed in the multiple canopy regions; determining the canopy density of each canopy region of the plant to be sprayed according to the leaf area index of each canopy region of the plant to be sprayed; inputting the canopy density of each canopy region of the plant to be sprayed into a correlation model to obtain the corresponding spray parameters of each canopy region of the plant to be sprayed; and the correlation model is used to represent the correlation between the canopy density and the spray parameters. The application realizes the accurate acquisition of the corresponding spray parameters of each canopy region, thereby improving the accuracy and efficiency of spraying.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of agricultural plant protection, and in particular to a spraying parameter acquisition method based on plant canopy density and a spraying machine. BACKGROUND

[0002] Fruit tree planting is an important part of modern agricultural production. With the use and popularization of pesticides, how to improve the spraying accuracy and efficiency, reduce waste and environmental pollution, etc. still needs to be solved.

[0003] The traditional fruit tree spraying method usually relies on manual experience to adjust the spraying amount and then sprays the whole tree. This method has many shortcomings. First, the spraying amount is usually large, which can easily lead to overuse of pesticides, increase costs, and cause unnecessary pollution to the environment. Second, due to differences in the growth state of different fruit trees, a uniform spraying amount cannot meet the actual needs of each tree, which can lead to insufficient spraying in some areas or excessive spraying in some areas, affecting the effectiveness of the pesticide, and even causing resource waste.

[0004] In recent years, variable rate spraying (VRS) technology based on crop growth characteristics has gradually become an effective means to solve the problems of traditional spraying. Variable rate spraying technology uses real-time sensing data, combines parameters such as crop growth state and vegetation coverage, and dynamically adjusts the spraying amount to achieve precise and efficient pesticide application.

[0005] However, existing variable rate spraying technology relies on static image recognition, sensor measurement, or simple algorithm models, resulting in the need for improved accuracy and efficiency of spraying. SUMMARY

[0006] The present application provides a spraying parameter acquisition method based on plant canopy density and a spraying machine to solve the problem of the need for improved accuracy and efficiency of spraying in the prior art, and to achieve accurate acquisition of spraying parameters, thereby improving the accuracy and efficiency of spraying.

[0007] The present application provides a spraying parameter acquisition method based on plant canopy density, comprising:

[0008] According to the crown point cloud data of the plant to be sprayed, the number of point clouds of the plant to be sprayed in multiple crown regions is determined;

[0009] According to the number of point clouds of the plant to be sprayed in multiple crown regions, the leaf area index of each crown region of the plant to be sprayed is determined;

[0010] According to the leaf area index of each crown region of the plant to be sprayed, the canopy density of each crown region of the plant to be sprayed is determined;

[0011] inputting the crown density of each crown layer region of the plant to be sprayed into the constructed correlation model, to obtain the corresponding spraying parameter of each crown layer region of the plant to be sprayed; the correlation model is used to represent the correlation between the crown density and the spraying parameter.

[0012] In some embodiments, the determining the point cloud quantity of each crown layer region of the plant to be sprayed according to the crown point cloud data of the plant to be sprayed comprises:

[0013] dividing the crown layer of the plant to be sprayed into a plurality of crown layer regions;

[0014] obtaining the point cloud data of the plurality of crown layer regions according to the crown point cloud data of the plant to be sprayed;

[0015] dividing the point cloud data of each crown layer region into a plurality of voxels, and calculating the point cloud quantity in each voxel;

[0016] determining the point cloud quantity of each crown layer region of the plant to be sprayed according to the point cloud quantity in each voxel and the plurality of voxels corresponding to each crown layer region.

[0017] In some embodiments, before the obtaining the point cloud data of the plurality of crown layer regions according to the crown point cloud data of the plant to be sprayed, the method further comprises:

[0018] clustering the crown point cloud data of the plant population based on the Euclidean distance clustering algorithm to obtain the clustering center point of each plant;

[0019] establishing a cylinder space model with the clustering center point of each plant as the center;

[0020] segmenting the crown point cloud data of the plant population by using the cylinder space model to obtain the crown point cloud data of each single plant.

[0021] In some embodiments, before the clustering the crown point cloud data of the plant population based on the Euclidean distance clustering algorithm to obtain the clustering center point of each plant, the method further comprises:

[0022] obtaining initial point cloud data of the vegetation crown layer from a plurality of different stations;

[0023] registering the initial point cloud data based on target ball fitting to obtain first point cloud data;

[0024] layering and simplifying the first point cloud data according to random sampling and curvature sampling to obtain second point cloud data;

[0025] filtering the second point cloud data based on HSI threshold and straight-through filtering to obtain the crown point cloud data of the plant population.

[0026] In some embodiments, the registration of the initial point cloud data based on the target ball fitting comprises:

[0027] detecting the target ball of the initial point cloud data and extracting the ball center coordinates;

[0028] matching the same target ball in the initial point cloud data of different stations based on the ball center coordinates and the ID of the target ball;

[0029] establishing the corresponding relationship between the initial point cloud data of different stations according to each matched target ball;

[0030] calculating the transformation matrix between the initial point cloud data of different stations according to the corresponding relationship;

[0031] registering the initial point cloud data based on the transformation matrix according to the ICP algorithm to obtain the first point cloud data.

[0032] In some embodiments, the hierarchical simplification of the first point cloud data according to random sampling and curvature sampling to obtain the second point cloud data comprises:

[0033] constructing a rectangular bounding box surrounding the first point cloud data;

[0034] dividing the spatial structure of the rectangular bounding box into multiple levels;

[0035] sampling the first point cloud data in each level by random sampling and curvature sampling to obtain the second point cloud data.

[0036] In some embodiments, the spray parameters include wind speed, travel speed and spray flow.

[0037] The application also provides a spraying machine, comprising: a moving wheel, a bearing plate, a fan, a liquid storage tank, a spray head, a data acquisition module, a laser radar and an electric control system.

[0038] The moving wheel is installed below the bearing plate, and the bearing plate bears the fan, the liquid storage tank, the laser radar and the electric control system.

[0039] The liquid storage tank is connected with the spray head, and the electric control system is connected with the laser radar, the spray head, the fan, the moving wheel and the data acquisition module respectively.

[0040] The moving wheel is used to drive the spraying machine to move.

[0041] The bearing plate is used to provide a bearing platform.

[0042] The fan is used to provide wind speed when spraying;

[0043] The liquid storage tank is used to store liquid to be sprayed;

[0044] The spray head is used to spray the liquid;

[0045] The data acquisition module is used to acquire the current rotating speed of the fan, the current driving speed of the mobile wheel, and the current spraying flow of the spray head;

[0046] The laser radar is used to acquire the crown point cloud data of the plant to be sprayed in the forward direction of the sprayer;

[0047] The electronic control system is used to execute the spraying parameter acquisition method based on the crown density of the plant according to the crown point cloud data of the plant to be sprayed, output the spraying parameter corresponding to each crown area of the plant to be sprayed, and automatically adjust the spraying flow of the spray head, the rotating speed of the fan, and the driving speed of the mobile wheel according to the spraying parameter corresponding to each crown area of the plant to be sprayed.

[0048] In some embodiments, the data acquisition module includes an encoder, a machine group sensor, and a flow sensor;

[0049] The encoder is nested in the connecting shaft between the fan and the motor, the machine group sensor is arranged beside the electronic control system, and the flow sensor is nested in the liquid delivery pipe of the spray head. The electronic control system is connected with the encoder, the machine group sensor, and the flow sensor respectively;

[0050] The motor is used to provide power for the fan;

[0051] The encoder is used to acquire the current rotating speed of the fan;

[0052] The machine group sensor is used to acquire the current driving speed of the mobile wheel;

[0053] The flow sensor is used to acquire the current spraying flow of the spray head.

[0054] In some embodiments, the electronic control system includes a computer, a flow controller, a wind speed controller, and a speed controller;

[0055] The flow controller is connected with the flow sensor, the spray head, and the computer respectively; the wind speed controller is connected with the encoder, the fan, and the computer respectively; and the speed controller is connected with the machine group sensor, the mobile wheel, and the computer respectively;

[0056] The computer is configured to execute the spray parameter acquisition method based on the plant canopy density according to the crown point cloud data of the plant to be sprayed, and output the spray parameter corresponding to each canopy region of the plant to be sprayed.

[0057] The flow controller is configured to automatically adjust the spray flow of the spray head according to the current spray flow of the spray head and the spray flow in the spray parameter corresponding to each canopy region of the plant to be sprayed.

[0058] The wind speed controller is configured to automatically adjust the rotation speed of the fan according to the current rotation speed of the fan and the wind speed in the spray parameter corresponding to each canopy region of the plant to be sprayed.

[0059] The speed controller is configured to automatically adjust the driving speed of the mobile wheel according to the current driving speed of the mobile wheel and the driving speed in the spray parameter corresponding to each canopy region of the plant to be sprayed.

[0060] The spray parameter acquisition method based on the plant canopy density and the sprayer provided by the application divide the canopy of the plant to be sprayed into a plurality of canopy regions, determine the leaf area index of each canopy region of the plant to be sprayed according to the point cloud quantity of the plant to be sprayed in the plurality of canopy regions, determine the canopy density of each canopy region of the plant to be sprayed according to the leaf area index of each canopy region of the plant to be sprayed, and acquire the spray parameter corresponding to each canopy region of the plant to be sprayed by constructing an association model for representing the association between the canopy density and the spray parameter. The application realizes accurate acquisition of the spray parameter corresponding to each canopy region by identifying the canopy density of each canopy region of the plant in real time and adjusting the spray parameter according to the canopy density of each canopy region, thereby improving the accuracy and efficiency of spraying, reducing the waste of pesticides, maximizing the effect of pesticide use, and reducing the environmental burden. BRIEF DESCRIPTION OF DRAWINGS

[0061] In order to more clearly illustrate the technical solutions in the application or prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the application, and other drawings can be obtained by those skilled in the art without any creative effort.

[0062] Figure 1 is one of the flowcharts of the spray parameter acquisition method based on the plant canopy density provided by the application.

[0063] Figure 2 is another flowchart of the spray parameter acquisition method based on the plant canopy density provided by the application.

[0064] Figure 3 Figure 3 is a flowchart of a method for obtaining spraying parameters based on plant canopy density according to the present application.

[0065] Figure 4 Figure 1 is a structural schematic diagram of a spraying machine according to the present application.

[0066] Figure 5 Figure 2 is a structural schematic diagram of a spraying machine according to the present application.

[0067] Figure 6 Figure 4 is a structural schematic diagram of an electric control system according to the present application.

[0068] Reference signs:

[0069] 1: moving wheel; 2: bearing plate; 3: fan;

[0070] 4: liquid storage tank; 5: spray head; 6: data acquisition module;

[0071] 61: encoder; 62: machine group sensor; 63: flow sensor;

[0072] 7: laser radar; 8: electric control system; 81: computer;

[0073] 82: flow controller; 83: wind speed controller; 84: speed controller;

[0074] 9: motor. DETAILED DESCRIPTION

[0075] In order to make the objectives, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be described clearly and completely below in conjunction with the drawings in the present application. Obviously, the described embodiments are some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0076] Figure 1 Figure 1 is a structural schematic diagram of a spraying machine according to the present application. Figure 1 A method for obtaining spraying parameters based on plant canopy density according to the present application comprises the following steps.

[0077] In step 110, the point cloud quantity of the plant to be sprayed in multiple canopy regions is determined according to the canopy point cloud data of the plant to be sprayed.

[0078] Specifically, the number of crown point clouds of the plant to be sprayed is acquired using a laser radar (LiDAR), and the crown of the plant to be sprayed is regionally divided to obtain the number of point clouds of the plant to be sprayed in multiple crown regions.

[0079] At step 120, the leaf area index of each crown region of the plant to be sprayed is determined according to the number of point clouds of the plant to be sprayed in multiple crown regions.

[0080] Specifically, for example, the crown of the plant to be sprayed is divided into an upper region, a middle region, a lower region, and a ground proximity region.

[0081] The upper region is located at the top of the crown and is usually exposed to strong sunlight. The leaf density here is relatively small, but due to sufficient light, it can be a high-incidence area of diseases and pests, especially during the mature period of crops, the reproduction of pests and the spread of diseases can first affect the upper part of the crown.

[0082] The middle region is located in the middle region of the crown and usually has a relatively uniform distribution of leaves. There are more leaves and moderate density here, and this part of the leaves usually accounts for the main part of the plant's absorption of sunlight and photosynthesis, and is also a potential transmission area of diseases and pests.

[0083] The lower region is located at the lower part of the crown, and the density of the leaves is usually high, but due to the proximity to the ground, the influence of light is small, and the air circulation is poor, which is easy to accumulate moisture and bacteria.

[0084] The ground proximity region is located at the bottom of the crown, close to the ground, with high air humidity. The leaves in the ground proximity region have low resistance to diseases and pests and are easily affected by underlying diseases and pests, which need special attention.

[0085] According to the number of point clouds of each crown region of the plant to be sprayed, the leaf area index of each crown region of the plant to be sprayed is determined. The specific formula is as follows:

[0086]

[0087] In the formula, LAI up represents the leaf area index of the upper region of the crown, P total represents the total number of point clouds of the entire crown, P up represents the number of point clouds of the upper region of the crown, LAI mid represents the leaf area index of the middle region of the crown, P mid represents the number of point clouds of the middle region of the crown, LAI low represents the leaf area index of the lower region of the crown P low represents the number of point clouds of the lower region of the crown, LAI ground represents the leaf area index of the ground proximity region, Pground The point cloud number represents the area close to the ground.

[0088] One data pair includes the point cloud number of each canopy area and the leaf area index, the point cloud number of the plurality of canopy areas can be obtained by using a laser radar or a sensor carried by a drone, and the corresponding leaf area index can be obtained at the same time and location by direct measurement or using remote sensing technology (for example, a ground spectral sensor, a drone, etc.).

[0089] In step 130, the canopy density of each canopy area of the plant to be sprayed is determined according to the leaf area index of each canopy area of the plant to be sprayed.

[0090] Specifically, the leaf area index is an important parameter for characterizing the canopy density of the plant, and the canopy density (Canopy Density) generally describes the density of the leaves of the plant canopy. According to the leaf area index of each canopy area of the plant to be sprayed, the canopy density of each canopy area of the plant to be sprayed can be determined.

[0091] The relationship between the leaf area index and the canopy density is as follows:

[0092] C = K x LAI

[0093] In the formula, C represents the canopy density, K is an empirical coefficient, which depends on the species of the plant, the type of the canopy, and the growth stage, etc., and LAI represents the leaf area index.

[0094] The canopy structure of different plant species is quite different, so a suitable coefficient K needs to be determined according to the canopy structure of a specific plant. For example, dense plants (such as forest trees) usually have a high k value, while sparse herbaceous plants (such as some crops) have a low K value.

[0095] In step 140, the canopy density of each canopy area of the plant to be sprayed is input into the constructed correlation model to obtain the corresponding spraying parameter of each canopy area of the plant to be sprayed; the correlation model is used to represent the correlation between the canopy density and the spraying parameter.

[0096] Specifically, the correlation model can output the optimal spraying parameter according to different canopy densities, and the correlation model can be obtained by a pesticide deposition test. In some embodiments, the spraying parameter can include: wind speed, driving speed and spraying flow.

[0097] For example, taking cotton as the spraying object, spraying flow, wind speed and spraying machine forward speed are selected as the experimental factors, and a plurality of levels are set for the cotton with different canopy densities to perform spraying deposition experiments, so as to obtain the pesticide utilization rate and deposition uniformity (coefficient of variation) of each group of experiments. According to the different emphasis on pursuing high-efficiency utilization of pesticides or uniform deposition of liquid medicine in the spraying requirement, the spraying target function is constructed by giving corresponding weights to the pesticide utilization rate and deposition uniformity. The Gaussian response surface is established according to the parameters and experimental results of the spraying deposition experiment, and then the optimal solution of each spraying target function is solved by using the genetic algorithm, so as to obtain the best correlation model corresponding to different canopy densities.

[0098] After obtaining the constructed correlation model, the canopy density of each canopy region of the plant to be sprayed is input into the constructed correlation model, so as to obtain the spraying parameters corresponding to each canopy region of the plant to be sprayed.

[0099] The spraying parameter acquisition method based on the canopy density of the plant provided in the embodiments of the present application divides the canopy of the plant to be sprayed into a plurality of canopy regions, determines the leaf area index of each canopy region of the plant to be sprayed according to the point cloud quantity of the plant to be sprayed in the plurality of canopy regions, determines the canopy density of each canopy region of the plant to be sprayed according to the leaf area index of each canopy region of the plant to be sprayed, and then acquires the spraying parameters corresponding to each canopy region of the plant to be sprayed by constructing a correlation model for representing the correlation between the canopy density and the spraying parameters. The present application realizes accurate acquisition of the spraying parameters corresponding to each canopy region by real-time identification of the canopy density of each canopy region of the plant and adjustment of the spraying parameters according to the canopy density of each canopy region, so as to improve the accuracy and efficiency of spraying, reduce the waste of pesticides, maximize the effect of pesticide use, and reduce the environmental burden.

[0100] In some embodiments, the point cloud quantity of the plant to be sprayed in the plurality of canopy regions is determined according to the canopy point cloud data of the plant to be sprayed, including:

[0101] The canopy of the plant to be sprayed is divided into a plurality of canopy regions;

[0102] The point cloud data of the plurality of canopy regions is acquired according to the canopy point cloud data of the plant to be sprayed;

[0103] The point cloud data of each canopy region is divided into a plurality of voxels, and the point cloud quantity in each voxel is calculated;

[0104] The point cloud quantity of the plant to be sprayed in the plurality of canopy regions is determined according to the point cloud quantity in each voxel and the plurality of voxels corresponding to each canopy region.

[0105] Specifically, the canopy of the plant to be sprayed is divided into multiple canopy regions, and the point cloud data of the canopy of the plant to be sprayed is divided into the multiple canopy regions to obtain point cloud data of the multiple canopy regions.

[0106] The point cloud data of each canopy region is divided into multiple voxels, and each voxel represents a spatial unit.

[0107] According to the multiple voxels corresponding to each canopy region, the number of point clouds in the multiple voxels is accumulated to obtain the number of point clouds of the plant to be sprayed in each canopy region.

[0108] The method for obtaining spray parameters based on the canopy density of a plant provided in the embodiments of the present application first divides the canopy of the plant to be sprayed into regions, reduces the point cloud space, then divides the point cloud data of each canopy region into multiple voxels to further reduce the point cloud space, and finally realizes accurate acquisition of the number of point clouds of the plant to be sprayed in each canopy region through the number of point clouds in each voxel.

[0109] In some embodiments, Figure 2 is a flowchart of the method for obtaining spray parameters based on the canopy density of a plant provided in the present application, as shown in Figure 2 Before obtaining the point cloud data of the multiple canopy regions according to the point cloud data of the canopy of the plant to be sprayed, the method further includes:

[0110] In step 210, the point cloud data of the canopy of the plant population is clustered based on the Euclidean distance clustering algorithm to obtain the cluster center point of each plant.

[0111] Specifically, the target point cloud data is clustered using the Euclidean distance clustering segmentation algorithm to extract the cluster center point of each plant. The clustering tolerance (for example, 0.02 meters, i.e., 2 centimeters) is set, which determines which points in the point cloud are considered part of the same cluster. The minimum and maximum cluster size are set to avoid small or large clusters affecting the results.

[0112] In step 220, a cylinder space model is established with the cluster center point of each plant as the center.

[0113] Specifically, a cylinder space model is established with the cluster center point of each plant as the center, i.e., the radius and height of the cylinder space model are set.

[0114] The radius of the cylinder space model is usually set according to the row spacing or plant size of the crop. For example, if the row spacing of the crop is fixed, the radius of the cylinder can be set to half of the row spacing to ensure that each cylinder space roughly corresponds to a plant in a row.

[0115] The height of the column space model is usually determined according to the growth height of the crop. The height of the plant varies at different growth stages, so the height of the column space model needs to be set according to the specific growth stage.

[0116] In step 230, the crown point cloud data of the plant population is segmented by using the column space model to obtain the crown point cloud data of each single plant.

[0117] Specifically, since each column space generally corresponds to a plant in a row, the crown point cloud data of each single plant can be obtained by segmenting the crown point cloud data of the plant population by using the column space model.

[0118] In some embodiments, the segmentation result can be verified to check the accuracy of the segmentation. In the case where the accuracy of the segmentation does not meet the standard, re-clustering and segmentation are performed.

[0119] The variable spraying method based on the crown density of the plant provided in the embodiments of the present application, by using the Euclidean distance clustering algorithm, the crown point cloud data of the plant population is clustered to obtain the clustering center point of each plant, and the column space model is established with the clustering center point of each plant as the center, the crown point cloud data of each single plant is obtained by segmenting the crown point cloud data of the plant population by using the column space model, and the crown point cloud data of each single plant is accurately obtained, and the subsequent crown point cloud data of the plant to be sprayed is processed.

[0120] In some embodiments, Figure 3 is a third flowchart of the spraying parameter acquisition method based on the crown density of the plant provided by the present application, as shown in Figure 3 Before the crown point cloud data of the plant population is clustered by using the Euclidean distance clustering algorithm to obtain the clustering center point of each plant, it further includes:

[0121] In step 310, the initial point cloud data of the vegetation canopy is obtained from multiple different stations.

[0122] Specifically, the initial point cloud data of the vegetation canopy is obtained by using a three-dimensional laser scanner (such as a ground laser radar scanning system). These data contain the phenotypic structure and spatial information of the vegetation, which provides a basis for subsequent density analysis.

[0123] In step 320, the initial point cloud data is registered based on target ball fitting to obtain the first point cloud data.

[0124] Specifically, the initial point cloud data collected from multiple different stations is registered to ensure the consistency and accuracy of the data. The point cloud registration method based on target ball fitting can be used to solve the ball center coordinates by fitting a spherical body to realize the registration of the point cloud data of different stations.

[0125] At step 330, the first point cloud data is hierarchically simplified according to random sampling and curvature sampling, to obtain second point cloud data.

[0126] Specifically, the point cloud data of the order of ten million is simplified to improve the work efficiency of subsequent processing and reduce the requirements for computer hardware and software. The hierarchical simplification method can be used to realize the point cloud data simplification in combination with random sampling and curvature sampling algorithms.

[0127] At step 340, the second point cloud data is filtered based on HSI threshold and straight-through filtering, to obtain the canopy point cloud data of the plant population.

[0128] Specifically, the collected RGB point cloud data is converted to the HSI color space, and appropriate HSI threshold is set according to the characteristics of the canopy of the crop population. Then, the second point cloud data is filtered by using the straight-through filtering in combination with the HSI color model, to automatically extract the canopy point cloud data of the plant population. This method improves the work efficiency and provides data support for the segmentation of the canopy point cloud data of the plant population.

[0129] The variable spraying method based on the canopy density of plants provided in the embodiments of the present application realizes automatic and efficient extraction of the canopy point cloud data of the plant population by performing fitting, simplification and filtering processing on the initial point cloud data of the vegetation canopy.

[0130] In some embodiments, the initial point cloud data is registered based on target ball fitting to obtain the first point cloud data, including:

[0131] The initial point cloud data is detected for target balls, and the ball center coordinates are extracted;

[0132] Based on the similarity of the ball center coordinates and the ID of the target balls, the same target ball is matched in the initial point cloud data of different stations;

[0133] According to each matched target ball, a corresponding relationship between the initial point cloud data of different stations is established;

[0134] According to the corresponding relationship, a transformation matrix between the initial point cloud data of different stations is calculated;

[0135] The initial point cloud data is registered based on the transformation matrix by using the ICP algorithm, to obtain the first point cloud data.

[0136] Specifically, the target balls are detected in the initial point cloud data by finding the point set of local maximum curvature, and the point set corresponds to the surface of the sphere. For each detected target ball, the sphere is fitted by using an algorithm (such as the RANSAC algorithm), and the ball center coordinates are extracted.

[0137] Target sphere matching is usually performed in the initial point cloud data of different stations based on the similarity of the sphere center coordinates and the same ID of the target spheres. For each matched target sphere, a mapping relationship between the initial point cloud data of different stations is established.

[0138] For example, assume there are station A and station B, a target sphere S A is detected in the point cloud data of station A, and its sphere center coordinate is C A = (x A , y A , z A ), and a matching target sphere S B is also detected in the point cloud data of station B, and its sphere center coordinate is C B = (x B , y B , z B ). Then this mapping relationship can be represented as That is, the mapping from the target sphere center coordinate in station A to the target sphere center coordinate in station B.

[0139] Using this mapping relationship, a transformation matrix between the initial point cloud data of different stations is calculated, that is, a transformation matrix (including translation and rotation) from the initial point cloud data of one station to the initial point cloud data of another station. Using the Iterative Closest Point (ICP) algorithm or other point cloud registration algorithms, the transformation matrix is continuously optimized through an iterative process, and the registration result of the initial point cloud data is further optimized according to the optimized transformation matrix, until a predetermined registration accuracy is reached, to obtain the first point cloud data.

[0140] The accuracy of the registration is verified by comparing the registered initial point cloud data with the known target sphere positions. The registration error is analyzed to ensure that the registration result meets the engineering or research requirements.

[0141] In some embodiments, before target sphere detection, target sphere selection and arrangement, data acquisition, and point cloud preprocessing are required, and the specific process is as follows:

[0142] Target sphere selection and arrangement: Select target spheres with high reflectivity to ensure clear identification in laser scanning. Arrange the target spheres within the field of view of each station to ensure that they can be detected in the data of multiple stations.

[0143] Data acquisition: Use a three-dimensional laser scanner to scan the target area while ensuring that the target spheres are included in the scanned data.

[0144] Point cloud preprocessing: The initial point cloud data collected is denoised to eliminate noise generated during the scanning process. Filtering algorithms such as radius filtering or statistical filtering are applied to further clean the point cloud data.

[0145] The method for obtaining spray parameters based on plant canopy density provided by the embodiments of the present application detects target balls from the initial point cloud data and extracts the ball center coordinates. Based on the similarity of the ball center coordinates and the ID of the target ball, the same target ball is matched in the initial point cloud data of different stations. According to each matched target ball, a corresponding relationship between the initial point cloud data of different stations is established. According to the corresponding relationship, a transformation matrix between the initial point cloud data of different stations is calculated. The initial point cloud data is registered based on the ICP algorithm to obtain the first point cloud data, thereby improving the registration accuracy.

[0146] In some embodiments, the first point cloud data is simplified by layering according to random sampling and curvature sampling to obtain the second point cloud data, including:

[0147] A rectangular bounding box enclosing the first point cloud data is constructed.

[0148] The spatial structure of the rectangular bounding box is divided into multiple levels.

[0149] Random sampling and curvature sampling are used to sample the first point cloud data in each level to obtain the second point cloud data.

[0150] Specifically, before constructing the rectangular bounding box, statistical filtering method is used to eliminate outliers and noise points, providing a clean data basis for subsequent simplification algorithms.

[0151] A rectangular bounding box (such as a cuboid bounding box or a cube bounding box) enclosing the first point cloud data is constructed, which can enclose all the first point cloud data.

[0152] The spatial structure of the rectangular bounding box is divided into multiple levels. In some embodiments, at each level, the point cloud space is further divided into smaller units, such as voxel units, to facilitate subsequent processing.

[0153] Random sampling is a simple and direct point cloud simplification method that reduces the amount of data by randomly selecting a certain number of point clouds. The curvature of each point cloud in the point cloud is calculated, which can provide important information about the local shape and features of the underlying surface. According to the local curvature value of the point cloud, points with larger curvature are preferentially retained, which usually correspond to feature points or edge points in the point cloud.

[0154] Random sampling and curvature sampling are applied at different levels to sample the first point cloud data to obtain second point cloud data, so as to achieve better simplification effect. In different levels, the proportion of random sampling and curvature sampling can be different. For example, the proportion of random sampling can be increased in flat areas, and the proportion of curvature sampling can be increased in complex or large curvature areas.

[0155] The method for obtaining spray parameters based on plant canopy density provided by the embodiment of the application achieves flexible sampling, adapts to different types of sampling areas, simplifies the number of point clouds, and improves the working efficiency of subsequent processing and reduces the requirements for computer software and hardware.

[0156] Figure 4 is one of the structural schematic diagrams of the sprayer provided by the application, Figure 5 is another structural schematic diagram of the sprayer provided by the application, as shown in Figure 4 and 5 The sprayer provided by the application comprises a moving wheel 1, a bearing plate 2, a fan 3, a liquid storage tank 4, a spray head 5, a data acquisition module 6, a laser radar 7, and an electric control system 8.

[0157] The moving wheel 1 is installed below the bearing plate 2, and the bearing plate 2 bears the fan 3, the liquid storage tank 4, the laser radar 7, and the electric control system 8;

[0158] The liquid storage tank 4 is connected with the spray head 5, and the electric control system 8 is connected with the laser radar 7, the spray head 5, the fan 3, the moving wheel 1, and the data acquisition module 6 respectively;

[0159] The moving wheel 1 is used to drive the sprayer to move;

[0160] The bearing plate 2 is used to provide a bearing platform;

[0161] The fan 3 is used to provide wind speed during spraying;

[0162] The liquid storage tank 4 is used to store liquid to be sprayed;

[0163] The spray head 5 is used to spray liquid;

[0164] The data acquisition module 6 is used to acquire the current rotating speed of the fan 3, the current driving speed of the moving wheel 1, and the current spray flow of the spray head 5;

[0165] The laser radar 7 is used to acquire canopy point cloud data of plants to be sprayed in the forward direction of the sprayer;

[0166] The electric control system 8 is used to execute any of the above-mentioned spray parameter acquisition methods based on the crown point cloud data of the plant to be sprayed, and output the corresponding spray parameters of each crown region of the plant to be sprayed; and automatically adjust the spray flow of the spray head 5, the speed of the fan 3 and the driving speed of the moving wheel 1 according to the corresponding spray parameters of each crown region of the plant to be sprayed.

[0167] Specifically, the spraying machine moves through the moving wheel 1, and the spraying machine comprises at least one moving wheel 1, for example, the number of moving wheels 1 can be 1, 2, 3 or 4.

[0168] The moving wheel 1 is installed below the bearing plate 2, and the bearing plate 2 provides a bearing platform, and the bearing plate 2 bears the fan 3, the liquid storage tank 4, the laser radar 7 and the electric control system 8. The fan 3 is used to provide the wind speed during spraying.

[0169] The liquid storage tank 4 is connected with the spray head 5, the liquid storage tank 4 stores the liquid to be sprayed, and the spray head 5 sprays the liquid stored in the liquid storage tank 4.

[0170] The electric control system 8 is connected with the laser radar 7, the spray head 5, the fan 3, the moving wheel 1 and the data acquisition module 6 respectively. The laser radar 7 collects the crown point cloud data of the plant to be sprayed in the forward direction of the spraying machine, and transmits the collected crown point cloud data to the electric control system 8. The electric control system 8 executes any of the above-mentioned spray parameter acquisition methods based on the crown density of the plant, and outputs the corresponding spray parameters of each crown region of the plant to be sprayed.

[0171] The data acquisition module 6 acquires the current speed of the fan 3, the current driving speed of the moving wheel 1 and the current spray flow of the spray head 5. The electric control system 8 obtains the current speed of the fan 3, the current driving speed of the moving wheel 1 and the current spray flow of the spray head 5 from the data acquisition module 6.

[0172] The electric control system 8 automatically adjusts the spray flow of the spray head 5, the speed of the fan 3 and the driving speed of the moving wheel 1 according to the current speed of the fan 3, the current driving speed of the moving wheel 1, the current spray flow of the spray head 5 and the corresponding spray parameters of each crown region of the plant to be sprayed, so that the adjusted wind speed, driving speed and spray flow meet the requirements of the corresponding spray parameters of each crown region of the plant to be sprayed.

[0173] The spraying machine provided by the embodiment of the application collects the crown point cloud data of the plants to be sprayed in the forward direction of the spraying machine through the laser radar 7, the data acquisition module 6 acquires the current rotating speed of the fan 3, the current driving speed of the mobile wheel 1 and the current spraying flow of the spray head 5, the electric control system 8 executes the spraying parameter acquisition method based on the crown density of the plants according to the crown point cloud data of the plants to be sprayed, and outputs the spraying parameter corresponding to each crown area of the plants to be sprayed, and then automatically adjusts the spraying flow of the spray head 5, the rotating speed of the fan 3 and the driving speed of the mobile wheel 1 according to the current rotating speed of the fan 3, the current driving speed of the mobile wheel 1, the current spraying flow of the spray head 5 and the spraying parameter corresponding to each crown area of the plants to be sprayed, so that the variable spraying is realized by adjusting the spraying parameter, and the spraying precision and efficiency are improved.

[0174] In some embodiments, the data acquisition module 6 comprises an encoder 61, a unit sensor 62 and a flow sensor 63.

[0175] The encoder 61 is nested in the connecting shaft between the fan 3 and the motor 9, the unit sensor 62 is arranged on the electric control system 8, and the flow sensor 63 is nested in the liquid conveying pipe of the spray head 5, and the electric control system 8 is connected with the encoder 61, the unit sensor 62 and the flow sensor 63 respectively.

[0176] The motor 9 is used to provide power for the fan 3.

[0177] The encoder 61 is used to acquire the current rotating speed of the fan 3.

[0178] The unit sensor 62 is used to acquire the current driving speed of the mobile wheel 1.

[0179] The flow sensor 63 is used to acquire the current spraying flow of the spray head 5.

[0180] Specifically, as shown in Figure 5 The data acquisition module 6 comprises the encoder 61, the unit sensor 62 and the flow sensor 63.

[0181] The encoder 61 is nested in the connecting shaft between the fan 3 and the motor 9, the motor 9 provides power for the fan 3, the encoder 61 acquires the current rotating speed of the fan 3, and transmits the acquired current rotating speed of the fan 3 to the electric control system 8.

[0182] The unit sensor 62 is arranged beside the electric control system 8, acquires the current driving speed of the mobile wheel 1, and transmits the acquired current driving speed of the mobile wheel 1 to the electric control system 8. For example, the unit sensor 62 is a GPS locator, the current driving distance and the current driving time of the spraying machine are acquired through the GPS locator, so that the current driving speed of the mobile wheel 1 is acquired.

[0183] The flow sensor 63 is nested in the delivery pipe of the spray head 5, collects the current spraying flow of the spray head 5, and transmits the collected current spraying flow of the spray head 5 to the electric control system 8.

[0184] The spraying machine provided by the embodiment of the present application collects the current rotating speed of the fan 3 through the encoder 61, collects the current driving speed of the mobile wheel 1 through the unit sensor 62, and collects the current spraying flow of the spray head 5 through the flow sensor 63, so as to realize real-time monitoring of the fan 3, the mobile wheel 1 and the spray head 5, and transmit the collected current rotating speed of the fan 3, the current driving speed of the mobile wheel 1 and the current spraying flow of the spray head 5 to the electric control system 8, so as to control the fan 3, the mobile wheel 1 and the spray head 5 by the electric control system 8.

[0185] In some embodiments, the electric control system 8 comprises a computer 81, a flow controller 82, a wind speed controller 83 and a speed controller 84.

[0186] The flow controller 82 is connected with the flow sensor 63, the spray head 5 and the computer 81 respectively; the wind speed controller 83 is connected with the encoder 61, the fan 3 and the computer 81 respectively; and the speed controller 84 is connected with the unit sensor 62, the mobile wheel 1 and the computer 81 respectively.

[0187] The computer 81 is used for executing any one of the above-mentioned spraying parameter acquisition methods based on the crown layer density of plants, and outputting the spraying parameter corresponding to each crown layer region of the plant to be sprayed.

[0188] The flow controller 82 is used for automatically adjusting the spraying flow of the spray head 5 according to the current spraying flow of the spray head 5 and the spraying flow in the spraying parameter corresponding to each crown layer region of the plant to be sprayed.

[0189] The wind speed controller 83 is used for automatically adjusting the rotating speed of the fan 3 according to the current rotating speed of the fan 3 and the wind speed in the spraying parameter corresponding to each crown layer region of the plant to be sprayed.

[0190] The speed controller 84 is used for automatically adjusting the driving speed of the mobile wheel 1 according to the current driving speed of the mobile wheel 1 and the driving speed in the spraying parameter corresponding to each crown layer region of the plant to be sprayed.

[0191] Specifically, Figure 6 is the structural diagram of the electric control system provided by the present application, as Figure 6 shown, the electric control system 8 comprises a computer 81, a flow controller 82, a wind speed controller 83 and a speed controller 84.

[0192] The flow controller 82 is connected with the flow sensor 63, the spray head 5 and the computer 81 respectively, and obtains the current spraying flow of the spray head 5 from the flow sensor 63 and obtains the spraying parameter corresponding to each crown layer area of the plant to be sprayed from the computer 81. The flow controller 82 compares the current spraying flow of the spray head 5 with the spraying flow in the spraying parameter corresponding to each crown layer area of the plant to be sprayed, and automatically adjusts the spraying flow of the spray head 5 according to the difference between the two, for example, by automatically adjusting the size and number of the spray holes of the spray head 5 to automatically adjust the spraying flow of the spray head 5, so that the current spraying flow of the adjusted spray head 5 is the spraying flow in the spraying parameter corresponding to each crown layer area of the plant to be sprayed, thereby realizing closed-loop feedback regulation of the spraying flow and ensuring the regulation accuracy of the spraying flow.

[0193] The wind speed controller 83 is connected with the encoder 61, the fan 3 and the computer 81 respectively, and obtains the current rotating speed of the fan 3 from the encoder 61 and obtains the spraying parameter corresponding to each crown layer area of the plant to be sprayed from the computer 81. The wind speed controller 83 determines the current wind speed according to the current rotating speed of the fan 3, compares the current wind speed with the wind speed in the spraying parameter corresponding to each crown layer area of the plant to be sprayed, and automatically adjusts the rotating speed of the fan 3 according to the difference between the two, for example, by indirectly adjusting the rotating speed of the fan 3 by adjusting the power provided by the motor 9, so that the current wind speed provided by the adjusted fan 3 is the wind speed in the spraying parameter corresponding to each crown layer area of the plant to be sprayed, thereby realizing closed-loop feedback regulation of the wind speed and ensuring the regulation accuracy of the wind speed.

[0194] The speed controller 84 is connected with the machine group sensor 62, the moving wheel 1 and the computer 81 respectively, and obtains the current driving speed of the moving wheel 1 from the machine group sensor 62 and obtains the spraying parameter corresponding to each crown layer area of the plant to be sprayed from the computer 81. The speed controller 84 compares the current driving speed of the moving wheel 1 with the driving speed in the spraying parameter corresponding to each crown layer area of the plant to be sprayed, and automatically adjusts the driving speed of the moving wheel 1 according to the difference between the two, so that the current driving speed of the adjusted moving wheel 1 is the driving speed in the spraying parameter corresponding to the plant to be sprayed, thereby realizing closed-loop feedback regulation of the driving speed and ensuring the regulation accuracy of the driving speed.

[0195] The spraying machine provided by the embodiment of the application automatically adjusts the spraying flow of the spray head 5 through the flow controller 82 according to the current spraying flow of the spray head 5 and the spraying flow in the spraying parameters corresponding to each crown layer area of the plant to be sprayed, automatically adjusts the rotating speed of the fan 3 through the wind speed controller 83 according to the current rotating speed of the fan 3 and the wind speed in the spraying parameters corresponding to each crown layer area of the plant to be sprayed, and automatically adjusts the driving speed of the moving wheel 1 through the speed controller 84 according to the current driving speed of the moving wheel 1 and the driving speed in the spraying parameters corresponding to each crown layer area of the plant to be sprayed, so that the closed-loop feedback regulation of the spraying parameters is realized, the regulation accuracy of the spraying parameters is ensured, and the spraying accuracy and efficiency are further improved.

[0196] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for obtaining spray parameters based on plant canopy density, characterized in that, The method comprises the following steps: determining the point cloud quantity of the plant to be sprayed in multiple canopy regions according to the canopy point cloud data of the plant to be sprayed; determining the leaf area index of each canopy region of the plant to be sprayed according to the point cloud quantity of the plant to be sprayed in multiple canopy regions; determining the canopy density of each canopy region of the plant to be sprayed according to the leaf area index of each canopy region of the plant to be sprayed; inputting the canopy density of each canopy region of the plant to be sprayed into a constructed correlation model to obtain the corresponding spraying parameters of each canopy region of the plant to be sprayed; the correlation model is used to represent the correlation between the canopy density and the spraying parameters; The method further comprises the following steps: clustering the canopy point cloud data of the plant population based on the Euclidean distance clustering algorithm to obtain the clustering center point of each plant; establishing a cylinder space model with the clustering center point of each plant as the center; segmenting the canopy point cloud data of the plant population by using the cylinder space model to obtain the canopy point cloud data of each single plant; Before clustering the canopy point cloud data of the plant population based on the Euclidean distance clustering algorithm to obtain the clustering center point of each plant, the method further comprises the following steps: obtaining initial point cloud data of the vegetation canopy from multiple different measuring stations; registering the initial point cloud data based on target ball fitting to obtain first point cloud data; simplifying the first point cloud data by layering based on random sampling and curvature sampling to obtain second point cloud data; filtering the second point cloud data based on HSI threshold and straight-through filtering to obtain the canopy point cloud data of the plant population; The method for registering the initial point cloud data based on target ball fitting to obtain first point cloud data comprises the following steps: detecting the initial point cloud data for target ball and extracting the ball center coordinates; matching the same target ball in the initial point cloud data of different measuring stations based on the ball center coordinates and the ID of the target ball; establishing a corresponding relationship between the initial point cloud data of different measuring stations according to each matched target ball; calculating the transformation matrix between the initial point cloud data of different measuring stations according to the corresponding relationship; registering the initial point cloud data based on the transformation matrix by using the ICP algorithm to obtain the first point cloud data. 2.The method of claim 1, wherein, The method for determining the point cloud quantity of the plant to be sprayed in multiple canopy regions according to the canopy point cloud data of the plant to be sprayed comprises the following steps: dividing the canopy of the plant to be sprayed into multiple canopy regions; obtaining the point cloud data of the multiple canopy regions according to the canopy point cloud data of the plant to be sprayed; dividing the point cloud data of each canopy region into multiple voxels and calculating the point cloud quantity in each voxel; determining the point cloud quantity of the plant to be sprayed in multiple canopy regions according to the point cloud quantity in each voxel and the multiple voxels corresponding to each canopy region. 3.The method of claim 1, wherein, The method for simplifying the first point cloud data by layering based on random sampling and curvature sampling to obtain second point cloud data comprises the following steps: constructing a rectangular bounding box surrounding the first point cloud data; dividing the spatial structure of the rectangular bounding box into multiple levels; The first point cloud data in each level is sampled by random sampling and curvature sampling to obtain the second point cloud data. 4.The method of claim 1, wherein, The spray parameters include wind speed, travel speed and spray flow.

5. A sprayer characterized by, It comprises: mobile wheels, a bearing plate, a fan, a liquid storage tank, a spray head, a data acquisition module, a laser radar and an electric control system; The mobile wheels are installed below the bearing plate, and the bearing plate bears the fan, the liquid storage tank, the laser radar and the electric control system; The liquid storage tank is connected with the spray head, and the electric control system is connected with the laser radar, the spray head, the fan, the mobile wheels and the data acquisition module respectively; The mobile wheels are used to drive the sprayer to move; The bearing plate is used to provide a bearing platform; The fan is used to provide wind speed during spraying; The liquid storage tank is used to store the liquid to be sprayed; The spray head is used to spray the liquid; The data acquisition module is used to acquire the current rotating speed of the fan, the current travel speed of the mobile wheels and the current spray flow of the spray head; The laser radar is used to acquire the crown point cloud data of the plant to be sprayed in the forward direction of the sprayer; The electric control system is used to execute the spray parameter acquisition method based on the crown density of the plant according to the crown point cloud data of the plant to be sprayed, output the spray parameters corresponding to each crown region of the plant to be sprayed, and automatically adjust the spray flow of the spray head, the rotating speed of the fan and the travel speed of the mobile wheels according to the spray parameters corresponding to each crown region of the plant to be sprayed.

6. The sprayer of claim 5, wherein, The data acquisition module comprises an encoder, a unit sensor and a flow sensor; The encoder is nested in the connecting shaft between the fan and the motor, the unit sensor is arranged beside the electric control system, and the flow sensor is nested in the liquid conveying pipe of the spray head. The motor is used to provide power for the fan; The encoder is used to acquire the current rotating speed of the fan; The unit sensor is used to acquire the current travel speed of the mobile wheels; The flow sensor is used to acquire the current spray flow of the spray head.

7. The sprayer of claim 6, wherein, The electric control system comprises a computer, a flow controller, a wind speed controller and a speed controller; The flow controller is connected with the flow sensor, the spray head and the computer respectively, the wind speed controller is connected with the encoder, the fan and the computer respectively, and the speed controller is connected with the unit sensor, the mobile wheels and the computer respectively; The computer is used to execute the spray parameter acquisition method based on the crown density of the plant according to the crown point cloud data of the plant to be sprayed, and output the spray parameters corresponding to each crown region of the plant to be sprayed; The flow controller is used to automatically adjust the spray flow of the spray head according to the current spray flow of the spray head and the spray flow in the spray parameters corresponding to each crown region of the plant to be sprayed. The wind speed controller is configured to automatically adjust the rotation speed of the fan according to the current rotation speed of the fan and the wind speed in the spray parameters corresponding to each canopy area of the plant to be sprayed. The speed controller is configured to automatically adjust the driving speed of the mobile wheel according to the current driving speed of the mobile wheel and the driving speed in the spray parameters corresponding to each canopy area of the plant to be sprayed.

Images