Intelligent management method and system for irrigation of landscaping plants

Abstract

The invention relates to the technical field of intelligent management of landscaping, in particular to an intelligent management method and system for irrigating landscaping plants, and the method comprises the following steps: 1, carrying out multi-dimensional data collection and dynamic modeling: collecting plant physiological data through a stem flow sensor, a leaf water potential sensor and a high-definition camera, the method comprises the following steps: acquiring environment and soil data through a meteorological station array and a soil sensor, and constructing a digital twinborn model containing a plant growth stage-water demand curve based on historical data; 2, water demand prediction and intelligent decision making: inputting real-time data into an LSTM neural network model to predict the water demand of the plant in the future 1-3 days, generating a dynamic irrigation scheme in combination with weather forecast data, and optimizing the irrigation amount and the time period according to the plant type, the growth stage and the soil partition; the water demand state is directly sensed through plant physiological signals, dynamic regulation and control are combined with a prediction model, water is saved by 50% or above compared with traditional regular irrigation, and the soil water accumulation rate is reduced to 5% or below.

Description

Technical Field

[0001] This invention relates to the field of intelligent management technology for landscaping, specifically to an intelligent management method and system for watering landscaping plants. Background Technology

[0002] The existing management of irrigation for landscaping and greening has significant technical limitations:

[0003] Extensive irrigation is inefficient: Traditional manual irrigation relies on experience and judgment, which leads to problems such as "over-irrigation" (water waste rate exceeding 40%) or "missed irrigation"; semi-automatic systems mostly adopt timed and quantitative modes, without considering the differences in plant species and growth stages. For example, pine and rose share the same irrigation scheme, resulting in more than 50% of plants being in a suboptimal water state.

[0004] The data perception dimension is limited: existing intelligent irrigation systems rely solely on soil moisture sensors to trigger irrigation, lacking direct monitoring of plant physiological states (such as leaf water potential and stem flow rate), and thus failing to reflect the plant's true water demand signals. For example, in the early stages of drought stress, soil moisture may not decrease significantly, but plants may already show signs of wilting.

[0005] Lack of predictive regulation: The system does not make forward-looking adjustments based on meteorological forecast data. If the system still irrigates according to the usual plan even if there is heavy rain the next day, it will cause waterlogging in the soil and root rot (occurrence rate of up to 25%). In addition, there is no growth cycle adaptation mechanism. For example, the water requirement of cherry blossoms during the flowering period is 3 times that during the dormant period, and the existing system cannot automatically adapt to this.

[0006] Insufficient system coordination: Irrigation execution equipment (valves, water pumps) and sensing equipment operate independently without closed-loop feedback. For example, if a nozzle is clogged and causes local water shortage, the system cannot automatically identify and adjust other nozzles to compensate, requiring manual troubleshooting (average time 3 hours / time). To address this, a smart management method and system for watering garden plants is proposed. Summary of the Invention

[0007] In view of this, the present invention provides a smart management method and system for watering garden plants, to solve or alleviate the technical problems existing in the prior art, and at least provide a beneficial option.

[0008] The technical solution of this invention is implemented as follows: A smart management method for watering garden plants, comprising the following steps:

[0009] Step 1: Multi-dimensional data acquisition and dynamic modeling: Plant physiological data are collected through stem flow sensors, leaf water potential sensors and high-definition cameras, and environmental and soil data are collected through weather station arrays and soil sensors. Based on historical data, a digital twin model containing the plant's "growth stage-water requirement curve" is constructed.

[0010] Step 2, Water Demand Prediction and Intelligent Decision-Making: Input real-time data into the LSTM neural network model to predict the water demand of plants in the next 1-3 days, and generate a dynamic irrigation plan by combining weather forecast data, optimizing the irrigation amount and time period according to plant type, growth stage and soil zoning.

[0011] Step 3, Adaptive Irrigation Execution: Achieve precise irrigation in zones through intelligent valve groups and variable frequency water pumps, use GPS and angle sensors to correct the nozzle range, and perform emergency handling for abnormal flow and extreme high temperature weather;

[0012] Step 4: Closed-loop feedback and model optimization: After irrigation, assess the water replenishment efficiency and plant health status, update the LSTM model parameters using reinforcement learning algorithms, and adapt new plant varieties through transfer learning.

[0013] More preferably, in step one, plant physiological data acquisition includes: stem flow sensor measurement accuracy ±0.1 L / h, leaf water potential sensor monitoring range -3.0~0 MPa, and high-definition camera image recognition extraction of wilting index (0-100%). Environmental and soil data acquisition includes: meteorological station array spacing 50-100 m, rainfall acquisition accuracy 0.1 mm, and solar radiation 0-2000 W / m². 2 The soil sensor is buried at a depth of 30cm and uploads volumetric water content (0-50%) and electrical conductivity data every 10 minutes. The "growth stage-water requirement curve" of the digital twin model includes: soil water content maintained at 25-30% during the seedling stage of trees and irrigation triggered when the stem flow rate is lower than 5L / h; soil water content of 20-25% and leaf water potential ≥-1.5MPa during the flowering period of flowering shrubs.

[0014] Further preferably, in step two, the prediction error of the LSTM neural network model is controlled within ±5%. Combined with the precipitation probability obtained from the meteorological department's API for the next 7 days, irrigation is automatically delayed when precipitation is ≥10mm within 24 hours. The differentiated optimization of the dynamic irrigation scheme includes: single irrigation amount for pine and cypress forests = 0.8 × (field water holding capacity - current water content) × soil area × bulk density; when the wilting index of rose bushes during flowering is >20%, the irrigation amount is increased by 15%. The irrigation time is preferentially selected from 4-6 am or 6-8 pm, and is executed when the wind speed is <3m / s.

[0015] Further preferably, in step three, the intelligent valve group supports the MODBUS protocol with a response time of <1s; the variable frequency water pump pressure adjustment range is 0.2-0.8MPa, the irrigation uniformity deviation is <5%, and the emergency handling includes: closing the valve and triggering a maintenance warning when the flow sensor detects a zone flow deviation of >20%; and starting the "cooling and water replenishment" mode in extreme high temperature (>35℃) to add one micro-spray with 30% of the normal water volume.

[0016] In a further preferred embodiment, the closed-loop feedback in step four includes: collecting soil moisture content and stem flow rate 1 hour after irrigation, generating a health report weekly based on leaf chlorophyll content and new shoot growth; model optimization using the PPO reinforcement learning algorithm, updating LSTM parameters monthly, and adding new plant varieties to adapt to the water requirement model through transfer learning.

[0017] A smart management system for watering garden plants includes the following modules:

[0018] Multi-source sensing module: includes plant physiological sensing unit (stem flow sensor, leaf water potential sensor, high-definition camera) and environmental and soil sensing unit (weather station array, soil sensor), used to collect plant physiological, environmental and soil data;

[0019] Intelligent decision-making module: including edge computing gateway and cloud decision engine. The edge computing gateway preprocesses data and stores high-frequency data from the last 72 hours. The cloud decision engine runs LSTM prediction model and reinforcement learning algorithm, and integrates plant knowledge base to generate dynamic irrigation plan.

[0020] Adaptive execution module: Composed of intelligent valve control unit (electric regulating valve, variable frequency water pump group) and mobile execution equipment (unmanned irrigation vehicle), it performs zoned irrigation according to the irrigation plan and handles abnormalities;

[0021] Human-computer interaction and operation and maintenance module: including a web management platform and a mobile APP, used to visualize data, receive anomaly alerts, support manual intervention and generate operation and maintenance reports.

[0022] Further preferably, in the plant physiological sensing unit: the stem flow sensor is a Dynamax SF-400 with a measurement range of 0-50 L / h; the leaf water potential sensor is a METERGroup WP4C with an accuracy of ±0.05 MPa; and the high-definition camera has a 1080P resolution and supports AI recognition of the wilting index. In the environmental and soil sensing unit: the weather station is a Davis Vantage Pro2 with a data update cycle of 1 minute; and the soil sensor is a Decagon 5TE that monitors water content, EC value, and temperature with an IP68 protection rating.

[0023] Preferably, the edge computing gateway uses an Advantech UNO-2484G industrial server, equipped with an ARM Cortex-A53 processor and a built-in InfluxDB real-time database; the cloud decision engine is deployed on an Alibaba Cloud ECS server, supporting concurrent processing of 100,000 data points and integrating a knowledge base containing 300+ plant water requirement parameters.

[0024] More preferably, the adaptive execution module includes: an intelligent valve control unit comprising a Belimo Z202 electric regulating valve (regulation accuracy ±1%) and a Grundfos CR15-6 variable frequency water pump set (flow rate 0-50m³ / h). 3 / h); The mobile execution device is an unmanned irrigation vehicle equipped with millimeter-wave radar obstacle avoidance, with a range of 8 hours.

[0025] In a further preferred embodiment, the human-computer interaction and operation and maintenance module includes: a web management platform with a data visualization dashboard (soil moisture heat map, plant health ranking), supporting manual intervention and operation log recording; a mobile APP that pushes abnormal warnings, supports voice command control, generates monthly water usage and health reports; a multi-source sensing module and an intelligent decision-making module that transmit data via LoRa / 4G; an intelligent decision-making module and an adaptive execution module that communicate via the MODBUS protocol; and a human-computer interaction module that connects to the cloud-based decision engine via an API interface.

[0026] The embodiments of the present invention have the following advantages due to the adoption of the above technical solutions:

[0027] I. This invention directly senses the water demand status through plant physiological signals and dynamically regulates it by combining a predictive model, saving more than 50% of water compared to traditional timed irrigation and reducing soil waterlogging rate to below 5%.

[0028] II. The differentiated solution of this invention is adapted to plant species and growth stages, reducing the leaf wilting index to 8%, increasing the annual growth of new shoots by 20%, and extending the flowering period of flowers by 15-20 days.

[0029] Third, the invention enables automatic detection and location of anomalies, reducing maintenance response time from 3 hours to 30 minutes; unmanned operation reduces labor costs by 60%, and annual maintenance costs are reduced by 450,000 yuan per 100,000 square meters of green space.

[0030] Fourth, the reinforcement learning model of this invention continuously optimizes the prediction accuracy, shortens the adaptation cycle of new plant varieties from 1 month to 1 week, and improves the system scalability by 3 times.

[0031] The above overview is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of the invention will become readily apparent from the accompanying drawings and the following detailed description. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 This is a flowchart of the method of the present invention;

[0034] Figure 2 This is a system flowchart of the present invention. Detailed Implementation

[0035] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0036] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0037] like Figure 1-2 As shown, this embodiment of the invention provides a smart management method for watering garden plants, including the following steps:

[0038] Step 1: Multi-dimensional data acquisition and dynamic modeling: Plant physiological data are collected through stem flow sensors, leaf water potential sensors and high-definition cameras, and environmental and soil data are collected through weather station arrays and soil sensors. Based on historical data, a digital twin model containing the plant's "growth stage-water requirement curve" is constructed.

[0039] Step 2, Water Demand Prediction and Intelligent Decision-Making: Input real-time data into the LSTM neural network model to predict the water demand of plants in the next 1-3 days, and generate a dynamic irrigation plan by combining weather forecast data, optimizing the irrigation amount and time period according to plant type, growth stage and soil zoning.

[0040] Step 3, Adaptive Irrigation Execution: Achieve precise irrigation in zones through intelligent valve groups and variable frequency water pumps, use GPS and angle sensors to correct the nozzle range, and perform emergency handling for abnormal flow and extreme high temperature weather;

[0041] Step 4: Closed-loop feedback and model optimization: After irrigation, assess the water replenishment efficiency and plant health status, update the LSTM model parameters using reinforcement learning algorithms, and adapt new plant varieties through transfer learning.

[0042] In one embodiment, in step one, plant physiological data acquisition includes: stem flow sensor measurement accuracy ±0.1 L / h, leaf water potential sensor monitoring range -3.0 to 0 MPa, and high-definition camera image recognition extraction of wilting index (0-100%). Environmental and soil data acquisition includes: meteorological station array spacing 50-100 m, rainfall acquisition accuracy 0.1 mm, and solar radiation 0-2000 W / m². 2 The soil sensor is buried at a depth of 30cm and uploads volumetric water content (0-50%) and electrical conductivity data every 10 minutes. The "growth stage-water requirement curve" of the digital twin model includes: soil water content maintained at 25-30% during the seedling stage of trees and irrigation triggered when the stem flow rate is lower than 5L / h; soil water content of 20-25% and leaf water potential ≥-1.5MPa during the flowering period of flowering shrubs.

[0043] In one embodiment, the prediction error of the LSTM neural network model in step two is controlled within ±5%. Combined with the meteorological department's API to obtain the precipitation probability for the next 7 days, irrigation is automatically delayed when the precipitation is ≥10mm within 24 hours. The differentiated optimization of the dynamic irrigation scheme includes: the single irrigation amount of pine and cypress forests = 0.8 × (field water holding capacity - current water content) × soil area × bulk density; when the wilting index of rose bushes during the flowering period is >20%, the irrigation amount is increased by 15%. The irrigation time is preferentially selected from 4-6 am or 18-20 pm, and it is executed when the wind speed is <3m / s.

[0044] In one embodiment, the intelligent valve group in step three supports the MODBUS protocol with a response time of <1s; the variable frequency water pump has a pressure regulation range of 0.2-0.8MPa, a water uniformity deviation of <5%, and emergency handling includes: closing the valve and triggering a maintenance warning when the flow sensor detects a zone flow deviation of >20%; and starting the "cooling and water replenishment" mode in extreme high temperatures (>35℃) to add one micro-spray of 30% of the normal water volume.

[0045] In one embodiment, the closed-loop feedback in step four includes: collecting soil moisture content and stem flow rate 1 hour after irrigation, generating a health report weekly based on leaf chlorophyll content and new shoot growth; model optimization using the PPO reinforcement learning algorithm, updating LSTM parameters monthly, and adding new plant varieties to adapt to the water requirement model through transfer learning.

[0046] A smart management system for watering garden plants includes the following modules:

[0047] Multi-source sensing module: includes plant physiological sensing unit (stem flow sensor, leaf water potential sensor, high-definition camera) and environmental and soil sensing unit (weather station array, soil sensor), used to collect plant physiological, environmental and soil data;

[0048] Intelligent decision-making module: including edge computing gateway and cloud decision engine. The edge computing gateway preprocesses data and stores high-frequency data from the last 72 hours. The cloud decision engine runs LSTM prediction model and reinforcement learning algorithm, and integrates plant knowledge base to generate dynamic irrigation plan.

[0049] Adaptive execution module: Composed of intelligent valve control unit (electric regulating valve, variable frequency water pump group) and mobile execution equipment (unmanned irrigation vehicle), it performs zoned irrigation according to the irrigation plan and handles abnormalities;

[0050] Human-computer interaction and operation and maintenance module: including a web management platform and a mobile APP, used to visualize data, receive anomaly alerts, support manual intervention and generate operation and maintenance reports.

[0051] In one embodiment, the plant physiological sensing unit includes: a stem flow sensor (Dynamax SF-400) with a measurement range of 0-50 L / h; a leaf water potential sensor (METERGroup WP4C) with an accuracy of ±0.05 MPa; and a high-definition camera with 1080P resolution that supports AI-based wilting index recognition. In the environmental and soil sensing unit, the weather station is a Davis Vantage Pro2 with a data update cycle of 1 minute; and the soil sensor is a Decagon 5TE that monitors water content, EC value, and temperature with an IP68 protection rating.

[0052] In one embodiment, the edge computing gateway uses an Advantech UNO-2484G industrial server equipped with an ARM Cortex-A53 processor and a built-in InfluxDB real-time database; the cloud decision engine is deployed on an Alibaba Cloud ECS server, supporting concurrent processing of 100,000 data points and integrating a knowledge base containing 300+ plant water requirement parameters.

[0053] In one embodiment, the adaptive execution module includes: a smart valve control unit comprising a Belimo Z202 electric regulating valve (regulation accuracy ±1%) and a Grundfos CR15-6 variable frequency water pump set (flow rate 0-50 m³ / h). 3 / h); The mobile execution device is an unmanned irrigation vehicle equipped with millimeter-wave radar obstacle avoidance, with a range of 8 hours.

[0054] In one embodiment, the human-computer interaction and operation and maintenance module includes: a web management platform with a data visualization dashboard (soil moisture heat map, plant health ranking), supporting manual intervention and operation log recording; a mobile APP pushes abnormal warnings, supports voice command control, generates monthly water usage and health reports; the multi-source sensing module and the intelligent decision-making module transmit data via LoRa / 4G; the intelligent decision-making module and the adaptive execution module communicate via the MODBUS protocol; and the human-computer interaction module connects to the cloud decision engine via an API interface.

[0055] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in the present invention, and these should all be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A smart management method for watering garden plants, characterized in that: Includes the following steps: Step 1: Multi-dimensional data acquisition and dynamic modeling: Plant physiological data are collected through stem flow sensors, leaf water potential sensors and high-definition cameras, and environmental and soil data are collected through weather station arrays and soil sensors. Based on historical data, a digital twin model containing the plant's "growth stage-water requirement curve" is constructed. Step 2, Water Demand Prediction and Intelligent Decision-Making: Input real-time data into the LSTM neural network model to predict the water demand of plants in the next 1-3 days, and generate a dynamic irrigation plan by combining weather forecast data, optimizing the irrigation amount and time period according to plant type, growth stage and soil zoning. Step 3, Adaptive Irrigation Execution: Achieve precise irrigation in zones through intelligent valve groups and variable frequency water pumps, use GPS and angle sensors to correct the nozzle range, and perform emergency handling for abnormal flow and extreme high temperature weather; Step 4: Closed-loop feedback and model optimization: After irrigation, assess the water replenishment efficiency and plant health status, update the LSTM model parameters using reinforcement learning algorithms, and adapt new plant varieties through transfer learning.

2. The intelligent management method for watering garden plants according to claim 1, characterized in that: In step one, plant physiological data acquisition includes: stem flow sensor measurement accuracy ±0.1 L / h, leaf water potential sensor monitoring range -3.0~0 MPa, and high-definition camera image recognition extraction of wilting index (0-100%). Environmental and soil data acquisition includes: meteorological station array spacing 50-100 m, rainfall acquisition accuracy 0.1 mm, and solar radiation 0-2000 W / m². 2 The soil sensor is buried at a depth of 30cm and uploads volumetric water content (0-50%) and electrical conductivity data every 10 minutes. The "growth stage-water requirement curve" of the digital twin model includes: soil water content maintained at 25-30% during the seedling stage of trees and irrigation triggered when the stem flow rate is lower than 5L / h; soil water content of 20-25% and leaf water potential ≥-1.5MPa during the flowering period of flowering shrubs.

3. The intelligent management method for watering garden plants according to claim 1, characterized in that: In step two, the prediction error of the LSTM neural network model is controlled within ±5%. Combined with the meteorological department's API to obtain the precipitation probability for the next 7 days, irrigation is automatically delayed when the precipitation is ≥10mm within 24 hours. The differentiated optimization of the dynamic irrigation scheme includes: the single irrigation amount of pine and cypress forests = 0.8 × (field water holding capacity - current water content) × soil area × bulk density; when the wilting index of rose bushes during the flowering period is >20%, the irrigation amount is increased by 15%; the irrigation time is preferentially selected from 4-6 am or 18-20 pm, and it is executed when the wind speed is <3m / s.

4. The intelligent management method for watering garden plants according to claim 1, characterized in that: The intelligent valve assembly in step three supports the MODBUS protocol and has a response time of <1s. The variable frequency water pump has a pressure regulation range of 0.2-0.8MPa and an irrigation uniformity deviation of <5%. Emergency handling includes: closing the valve and triggering a maintenance warning when the flow sensor detects a zone flow deviation of >20%; and starting the "cooling and water replenishment" mode in extreme high temperature (>35℃) to add one micro-spray with 30% of the normal water volume.

5. The intelligent management method for watering garden plants according to claim 1, characterized in that: The closed-loop feedback in step four includes: collecting soil moisture content and stem flow rate 1 hour after irrigation, generating a health report weekly based on leaf chlorophyll content and new shoot growth; model optimization using the PPO reinforcement learning algorithm, updating LSTM parameters monthly, and adding new plant varieties to adapt to the water requirement model through transfer learning.

6. A smart management system for watering garden plants, coupled with a smart management method for watering garden plants as described in any one of claims 1-5, characterized in that: Includes the following modules: Multi-source sensing module: includes plant physiological sensing unit (stem flow sensor, leaf water potential sensor, high-definition camera) and environmental and soil sensing unit (weather station array, soil sensor), used to collect plant physiological, environmental and soil data; Intelligent decision-making module: It includes an edge computing gateway and a cloud decision engine. The edge computing gateway preprocesses data and stores high-frequency data from the last 72 hours. The cloud decision engine runs an LSTM prediction model and reinforcement learning algorithm, and integrates a plant knowledge base to generate dynamic irrigation plans. Adaptive execution module: Composed of intelligent valve control unit (electric regulating valve, variable frequency water pump group) and mobile execution equipment (unmanned irrigation vehicle), it performs zoned irrigation according to the irrigation plan and handles abnormalities; Human-computer interaction and operation and maintenance module: including a web management platform and a mobile APP, used to visualize data, receive anomaly alerts, support manual intervention and generate operation and maintenance reports.

7. A smart management system for watering garden plants according to claim 6, characterized in that: The plant physiological sensing unit includes: a stem flow sensor (Dynamax SF-400, measuring range 0-50 L / h); a leaf water potential sensor (METERGroup WP4C, accuracy ±0.05 MPa); and a high-definition camera with 1080P resolution, supporting AI recognition of wilting index. The environmental and soil sensing unit includes: a weather station (Davis Vantage Pro2, data update cycle 1 minute); and a soil sensor (Decagon 5TE, monitoring water content, EC value, and temperature, with an IP68 protection rating).

8. A smart management system for watering garden plants according to claim 6, characterized in that: The edge computing gateway uses Advantech UNO-2484G industrial server, equipped with ARM Cortex-A53 processor and built-in InfluxDB real-time database; the cloud decision engine is deployed on Alibaba Cloud ECS server, supports 100,000-level concurrent data processing, and integrates a knowledge base containing 300+ plant water requirement parameters.

9. A smart management system for watering garden plants according to claim 6, characterized in that: The adaptive execution module includes: an intelligent valve control unit comprising a Belimo Z202 electric regulating valve (regulation accuracy ±1%) and a Grundfos CR15-6 variable frequency water pump set (flow rate 0-50m³ / h). 3 / h); The mobile execution device is an unmanned irrigation vehicle equipped with millimeter-wave radar obstacle avoidance, with a range of 8 hours.

10. A smart management system for watering garden plants according to claim 6, characterized in that: The human-computer interaction and operation and maintenance module includes: a web management platform with a data visualization dashboard (soil moisture heat map, plant health ranking), supporting manual intervention and operation log recording; a mobile APP that pushes abnormal warnings, supports voice command control, and generates monthly water usage and health reports; a multi-source sensing module and an intelligent decision-making module that transmit data via LoRa / 4G; an intelligent decision-making module and an adaptive execution module that communicate via the MODBUS protocol; and a human-computer interaction module that connects to the cloud decision engine via an API interface.

Images