Search

CN-122024099-A - Urban garden management system and method based on artificial intelligence

CN122024099ACN 122024099 ACN122024099 ACN 122024099ACN-122024099-A

Abstract

The invention relates to the field of garden management, in particular to an artificial intelligence-based urban garden management system and method, comprising the following steps: the system comprises a vegetation mapping module, a garden evaluation module, a transpiration water-requiring module, an irrigation decision module and a planning control module, wherein the vegetation mapping module is used for building a vegetation model, the garden evaluation module is used for estimating the three-dimensional green amount of gardens, the transpiration water-requiring module is used for predicting the vegetation water-requiring amount, the irrigation decision module is used for building a water-requiring amount model of a garden, and the planning control module is used for building a flow control system for irrigation.

Inventors

  • XU PENG
  • YAO LIBAO
  • ZHU YEHUA

Assignees

  • 江苏百绿环境科技有限公司

Dates

Publication Date
20260512
Application Date
20260202

Claims (10)

  1. 1. An artificial intelligence-based urban garden management method, which is characterized by comprising the following steps: S1, shooting a garden area by using an unmanned aerial vehicle, generating a time sequence point cloud according to vegetation image flows in a vegetation shooting frame, calculating vegetation branch radius according to point cloud data, clustering and dividing the point cloud according to a framework, generating a quadrilateral blade structure at edge nodes of the framework, and generating a vegetation model; S2, extracting a vegetation leaf area index from the image, determining vegetation coverage of a garden area according to the vegetation color spliced image, performing voxel segmentation on a vegetation model, and estimating a three-dimensional green amount of a garden according to the vegetation coverage; S3, collecting soil index data, screening a minimum data set, constructing a membership function, calculating a soil quality index, calculating a water stress index according to the soil quality index, a leaf area index and a vegetation type, and predicting vegetation water demand according to the water stress index and the three-dimensional green amount; S4, training a neural network through meteorological data, establishing a fuzzy decision system, inputting precipitation, humidity and temperature difference, outputting meteorological evaporation capacity, establishing a water demand model according to vegetation water demand, meteorological evaporation capacity and soil water content, and outputting water demand for irrigation in a park; s5, establishing a prediction model of soil moisture content, determining irrigation quantity according to water demand and the soil moisture content, dividing gardens into subareas according to a central support shaft and an irrigation radius of an irrigation device, establishing a flow control system, and adjusting flow, irrigation time and irrigation interval in each subarea.
  2. 2. The urban garden management method based on artificial intelligence according to claim 1, wherein the step S1 comprises the following steps: S11, planning a gridding route of the unmanned aerial vehicle in a garden area, checking color indexes of a shot image, generating a shooting frame, matching the position of the shooting frame according to SIFT features of the height change rate of the unmanned aerial vehicle, enabling the shooting frame to contain complete vegetation data, and matching and compensating the image according to the speed change rate; S12, registering images in different time to the same coordinate system according to the depth images and the time stamps shot by the unmanned aerial vehicle, sampling and processing the depth image frames in the coordinate system to generate a single plant vegetation point cloud, and extracting branch initial skeleton lines through the Voronoi diagram; S13, weighting and distributing weights according to the length, the point cloud density and the average curvature of each branch skeleton in proportion, removing low-weight skeletons, integrating adjacent skeletons, performing cylindrical fitting along skeleton lines, generating quadrilateral blades with the same size at end nodes and middle nodes of the skeletons, and determining the blade area by vegetation types.
  3. 3. The urban garden management method based on artificial intelligence according to claim 2, wherein the step S2 comprises: S21, calculating an NDVI index of vegetation through a shot spectral image, inputting a radiation transmission model to obtain a leaf area index, separating ground points and vegetation point pixels, splicing all the vegetation point pixels, obtaining a vegetation binary image through semantic segmentation, and calculating the proportion of vegetation pixels to total pixels to obtain vegetation coverage; s22, voxelizing a three-dimensional vegetation model, calculating the surface area of a blade, the surface area of a branch and the volume of vegetation in each voxel, and summarizing the sum of all the surface areas of all the vegetation to obtain the three-dimensional green amount of the urban garden area; The step S3 comprises the following steps: S31, synchronizing soil index data from a meteorological station or a soil monitoring station, wherein the soil index data comprise conductivity, carbon nitrogen ratio, sand content and water content, time-space alignment data, analyzing and screening indexes by using main components to form a minimum data set, constructing an S-type membership function for each index, and calculating a soil quality index; S32, fitting a water stress index of each vegetation unit based on the vegetation pore air permeability model, calculating potential evapotranspiration by using a Pengman formula, training a prediction model by using a long-period memory network, inputting historical meteorological data, soil water content, vegetation index and potential evapotranspiration, and outputting vegetation water demand in one period in the future.
  4. 4. The method for managing urban gardens based on artificial intelligence according to claim 3, wherein the step S4 comprises the following steps: S41, taking each meteorological variable as an input layer node of a neural network, taking the random combination of the meteorological variables as a training data set, determining the number of hidden layer nodes of the neural network, and optimizing the connection weight to obtain a trained BP neural model; and S42, complementarily weighting and combining the precipitation, irrigation, vegetation water demand and runoff, as a prediction model of the soil moisture content, adjusting irrigation parameters in the prediction model, taking the irrigation as a target, and maintaining the soil moisture content in the next period to be kept in a preset range, so as to generate an irrigation scheduling scheme.
  5. 5. The method for managing urban gardens based on artificial intelligence according to claim 4, wherein the step S5 comprises the following steps: S51, predicting soil moisture content through a radial basis function neural network according to urban weather characteristics including seasons, temperatures, precipitation and humidity, adjusting center vectors and width parameters by using a gradient descent method, and determining availability of water demand; s52, dividing subareas by taking coordinates of an irrigation center fulcrum as a circle center and effective ranges as radiuses, setting regional irrigation coefficients in each subarea according to terrains and soil moisture content, adjusting irrigation time of irrigation spray head equipment of each subarea according to the irrigation coefficients and an irrigation scheduling scheme, And S53, controlling the switch of the electromagnetic valve through the controller, adjusting the frequency of the pump station and the opening of the valve, irrigating according to the start-stop time sequence of the spray heads in each subarea, and displaying the irrigation time and the water consumption on the interactive interface.
  6. 6. An artificial intelligence-based urban garden management system is characterized by comprising a vegetation mapping module, a garden evaluation module, a transpiration water demand module, an irrigation decision module and a planning control module; The vegetation mapping module is used for shooting urban garden areas by using unmanned aerial vehicles, acquiring a vegetation shooting frame through pre-detection of picture frames, generating time sequence point cloud data according to target vegetation image streams in the shooting frame, carrying out centralized processing on the point cloud data, calculating vegetation branch radius, clustering and dividing the point cloud according to a framework, simulating convergence effect according to a different-speed growth algorithm, and generating a quadrilateral blade model at edge nodes of the framework to obtain a vegetation model; The garden evaluation module is used for extracting vegetation leaf area indexes from unmanned aerial vehicle images, removing soil background by utilizing vegetation canopy chromatic aberration, splicing images, extracting canopy data according to image gray scale, determining the height of a garden area canopy and vegetation coverage, and estimating the three-dimensional green amount of a garden by voxel segmentation of a vegetation model; The transpiration water demand module is used for collecting soil index data, screening a minimum data set through principal component analysis, calculating a soil quality index according to membership functions and weights, calculating a water stress index according to the soil quality index, leaf area index and vegetation types through a plant growth model, inputting the water stress index and three-dimensional green quantity into a Pengman formula, calculating vegetation transpiration quantity, and predicting vegetation water demand by using a long-short-period memory neural network; the irrigation decision module is used for training the BP neural network through meteorological data, establishing a fuzzy decision system through the trained model, outputting meteorological evaporation capacity by taking precipitation, humidity and temperature difference as input quantities, establishing a water demand model of the park according to vegetation water demand, meteorological evaporation capacity and soil moisture content, and evaluating water demand of park irrigation to enable the soil moisture content in the park to be kept in a preset range; The planning control module is used for establishing an RBF prediction model of soil moisture content, determining irrigation quantity according to water demand and the soil moisture content, dividing gardens into fan-shaped subareas with different radius and other central angles according to the central pivot coordinates and the irrigation radius of the irrigation device, taking the minimum irrigation time difference of equipment in each subarea as a target, taking flow, irrigation time and irrigation interval as constraints, and establishing a flow control system for irrigation.
  7. 7. The urban garden management system based on artificial intelligence according to claim 6, wherein the vegetation mapping module comprises an image acquisition unit, a point cloud processing unit and a branch modeling unit; The image acquisition unit is used for planning a gridding route of the unmanned aerial vehicle in a garden area, checking color indexes of a shot image, generating a shooting frame, matching the position of the shooting frame according to SIFT features of the height change rate of the unmanned aerial vehicle, enabling the shooting frame to contain complete vegetation data, and matching and compensating the image according to the speed change rate; The point cloud processing unit is used for registering images in different time to the same coordinate system according to the depth images and the time stamps shot by the unmanned aerial vehicle, sampling and processing the depth image frames in the coordinate system to generate single plant vegetation point clouds, and extracting branch initial skeleton lines through the Voronoi diagram; The branch modeling unit is used for removing low-weight frameworks according to the weighted sum distribution weight of the length, the point cloud density and the average curvature of each branch framework in proportion, integrating adjacent frameworks, performing cylinder fitting along skeleton lines, generating quadrilateral blades with the same size at end nodes and middle nodes of the frameworks, and determining the blade area by vegetation types.
  8. 8. The urban garden management system based on artificial intelligence according to claim 7, wherein the garden evaluation module comprises an image stitching unit and a greening monitoring unit; the image splicing unit is used for calculating the NDVI index of vegetation through the shot spectral images, inputting a radiation transmission model to obtain a leaf area index, separating ground points and vegetation point pixels, splicing all the vegetation point pixels, obtaining a vegetation binary image through semantic segmentation, and calculating the proportion of vegetation pixels to total pixels to obtain vegetation coverage; the greening monitoring unit is used for voxelizing the three-dimensional vegetation model, calculating the surface area of the blade, the surface area of the branch and the volume of the vegetation in each voxel, and summarizing the sum of all the surface areas of all the vegetation to obtain the three-dimensional green amount of the urban garden area; the transpiration water demand module comprises a soil quality unit and a demand prediction unit; The soil quality unit is used for synchronizing soil index data from a meteorological station or a soil monitoring station, the soil index data comprise conductivity, carbon nitrogen ratio, sand content and water content, time-space alignment data, main component analysis and screening indexes are used for forming a minimum data set, an S-shaped membership function is constructed for each index, and a soil quality index is calculated; The demand prediction unit is used for fitting the water stress index of each vegetation unit based on the vegetation pore air permeability model, calculating potential evapotranspiration by using a Pengman formula, training the prediction model by using a long-period memory network, inputting historical meteorological data, soil water content, vegetation index and potential evapotranspiration, and outputting vegetation water demand in one period in the future.
  9. 9. The urban garden management system based on artificial intelligence according to claim 8, wherein the irrigation decision module comprises a model training unit and a water use evaluation unit; The model training unit is used for taking each meteorological variable as an input layer node of the neural network, randomly combining the meteorological variables as a training data set, determining the number of hidden layer nodes of the neural network, optimizing the connection weight, and obtaining a trained BP neural model; the water consumption evaluation unit is used for carrying out supplementary weighted combination on precipitation, irrigation, vegetation water demand and runoff, taking the supplementary weighted combination as a prediction model of soil moisture content, adjusting irrigation parameters in the supplementary weighted combination, taking the minimum irrigation water content as a target, maintaining the soil moisture content in the next period to be in a preset range, and generating an irrigation scheduling scheme.
  10. 10. The urban garden management system based on artificial intelligence according to claim 9, wherein the planning control module comprises a soil moisture content detection unit, a regional irrigation unit and a flow control unit; the soil moisture content detection unit is used for predicting soil moisture content through a radial basis function neural network according to urban weather characteristics including seasons, temperature, precipitation and humidity, adjusting center vectors and width parameters by using a gradient descent method, and determining the availability of water demand; The regional irrigation unit is used for dividing subareas by taking the coordinates of an irrigation center pivot as a circle center and the effective range as a radius, setting regional irrigation coefficients in each subarea according to the topography and soil moisture content, adjusting the irrigation time of the irrigation spray head equipment of each subarea according to the irrigation coefficients and the irrigation scheduling scheme, The flow control unit is used for controlling the switch of the electromagnetic valve through the controller, adjusting the frequency of the pump station and the opening of the valve, irrigating according to the start-stop time sequence of the spray heads in each subarea, and displaying the irrigation time and the water consumption on the interactive interface.

Description

Urban garden management system and method based on artificial intelligence Technical Field The invention relates to the field of garden management, in particular to an artificial intelligence-based urban garden management system and method. Background Urban gardens are a kind of natural space which is formed by plants, terrains and buildings in urban areas, and can provide leisure, ornamental and ecological environment improvement functions as an important component of urban ecology and public services. Along with the development of urban management level, the automation level of urban garden construction is continuously improved, and in order to solve the problem of continuously increasing water demand in the green land irrigation process, the intelligent irrigation system formed by water-saving irrigation equipment such as sprinkling irrigation and drip irrigation is utilized for accurately supplying water, so that the intelligent irrigation system becomes a main stream irrigation mode of urban gardens. The management process of urban gardens relies on loaded down with trivial details on-site measurement, and the process work load of measuring and calculating regional vegetation in gardens is big, and the aassessment to vegetation irrigation process is inaccurate, and in gardens, vegetation distribution heterogeneity has also increased the measurement and calculation degree of difficulty. The existing automatic irrigation device has only a single water quantity adjusting function, is difficult to cope with irrigation requirements of different vegetation in space-time variation, reduces the utilization efficiency of water, causes waste of irrigation water, and also easily affects the growth process of garden plants. In addition, vegetation irrigation quantity is closely related to vegetation growth, soil and climate, but under actual measurement conditions, the soil structure is different, irrigation quantity results are easily interfered by matrix effects, irrigation accuracy is reduced, the cost of a crop sensing system is higher, sensors are difficult to lay, and the problems of high cost, poor timeliness, large error and the like exist. Disclosure of Invention The invention aims to provide an artificial intelligence-based urban garden management system and method, which are used for solving the problems in the background technology. In order to solve the technical problems, the invention provides the technical scheme that the urban garden management system based on artificial intelligence comprises a vegetation mapping module, a garden evaluation module, a transpiration water demand module, an irrigation decision module and a planning control module; The vegetation mapping module is used for shooting urban garden areas by using unmanned aerial vehicles, acquiring a vegetation shooting frame through pre-detection of picture frames, generating time sequence point cloud data according to target vegetation image streams in the shooting frame, carrying out centralized processing on the point cloud data, calculating vegetation branch radius, clustering and dividing the point cloud according to a framework, simulating convergence effect according to a different-speed growth algorithm, and generating a quadrilateral blade model at edge nodes of the framework to obtain a vegetation model; The garden evaluation module is used for extracting vegetation leaf area indexes from unmanned aerial vehicle images, removing soil background by utilizing vegetation canopy chromatic aberration, splicing images, extracting canopy data according to image gray scale, determining the height of a garden area canopy and vegetation coverage, and estimating the three-dimensional green amount of a garden by voxel segmentation of a vegetation model; The transpiration water demand module is used for collecting soil index data, screening a minimum data set through principal component analysis, calculating a soil quality index according to membership functions and weights, calculating a water stress index according to the soil quality index, leaf area index and vegetation types through a plant growth model, inputting the water stress index and three-dimensional green quantity into a Pengman formula, calculating vegetation transpiration quantity, and predicting vegetation water demand by using a long-short-period memory neural network; the irrigation decision module is used for training the BP neural network through meteorological data, establishing a fuzzy decision system through the trained model, outputting meteorological evaporation capacity by taking precipitation, humidity and temperature difference as input quantities, establishing a water demand model of the park according to vegetation water demand, meteorological evaporation capacity and soil moisture content, and evaluating water demand of park irrigation to enable the soil moisture content in the park to be kept in a preset range; The planning control module is used for establ