An automated smart flowerpot with water volume monitoring and early warning functions

Abstract

This invention discloses an automated smart flowerpot with water level monitoring and early warning functions, belonging to the field of smart flowerpot technology. It includes: a flowerpot body, designed as an integrated unit, combining water storage, planting, and irrigation functions; a smart controller, installed inside the flowerpot body, integrating environmental data acquisition, irrigation logic judgment, early warning triggering, and communication interaction functions, collecting soil moisture and water level data, executing control commands and performing negative feedback adjustment and verification, supporting multi-level water level early warning reminders; and a mobile management terminal, maintaining wireless communication with the smart controller, visualizing environmental monitoring data, providing suitable plant recommendations and plant type and growth status identification, and generating personalized maintenance strategies using a Q-learning algorithm. This invention achieves the integration of water storage, planting, and irrigation functions through an integrated layered design; and improves maintenance efficiency and portability through a dual modular design of mechanical and electrical controls.

Description

Technical Field

[0001] This invention relates to the field of smart flowerpot technology, and in particular to an automated smart flowerpot with water volume monitoring and early warning functions. Background Technology

[0002] Traditional flowerpots, as the basic containers for plant cultivation, have long relied on manual experience for water management and environmental regulation, resulting in low maintenance efficiency and significant resource waste. Ordinary terracotta or plastic pots lack environmental sensing capabilities, requiring users to frequently check soil moisture manually. This can easily lead to plant death due to untimely or excessive watering (for example, the mortality rate of green plants caused by forgetting to water during work hours can exceed 30%). With accelerating urbanization and shrinking living spaces, people's demand for convenient plant care is increasingly urgent, but the functional limitations of traditional flowerpots make it difficult to meet the pace of modern life.

[0003] Smart flowerpots overcome the aforementioned limitations by integrating sensors and IoT technology. Their core functionality involves embedding soil moisture, water level, temperature, and light sensors within the flowerpot body to collect real-time data on the plant's growth environment. Current smart flowerpot technology primarily focuses on three core functions: environmental monitoring, automated irrigation, and user interaction. In terms of environmental monitoring, mainstream products generally use basic sensors (such as the YL-69 soil moisture sensor, water level sensor, and DS18B20 temperature and humidity sensor) to collect data and synchronize it to a mobile app via IoT technology (Wi-Fi / NB-IoT). Users can remotely view plant status and trigger irrigation commands. Automated irrigation systems often rely on preset thresholds to control water pumps or solenoid valves. For example, a miniature submersible pump automatically starts when soil moisture falls below a set value, combined with drip irrigation technology for precise water supply. Regarding user interaction, some products integrate LED indicator lights, local alarms with buzzers, or push notifications via an app; a few high-end models support voice interaction and dual-screen display of environmental data.

[0004] Current smart flowerpot technology faces several technical challenges: Firstly, the accuracy of environmental sensing is insufficient. Soil moisture sensors are easily affected by salinity, leading to reading drift, and a single sensor cannot comprehensively reflect the plant's water potential, increasing the risk of incorrect irrigation. Secondly, the adaptability of intelligent decision-making is limited. Current systems largely rely on preset threshold rules and lack the ability to dynamically learn from the plant's growth stages. Although some research has attempted to introduce machine learning algorithms to optimize irrigation strategies, the individual differences among potted plants result in low accuracy of general models. Summary of the Invention

[0005] Therefore, it is necessary to provide an automated smart flowerpot with water volume monitoring and early warning functions to address the aforementioned technical problems.

[0006] This invention provides an automated smart flowerpot with water level monitoring and early warning functions, comprising:

[0007] The main body of the flowerpot is designed in an integrated manner, combining water storage, planting and irrigation functions;

[0008] The intelligent controller is set on the end face of the flowerpot body and is used to integrate environmental data acquisition, irrigation logic judgment, early warning triggering and communication interaction functions. It collects soil moisture and water level data, executes control commands and performs negative feedback adjustment and verification, and supports multi-level water volume early warning reminders.

[0009] The mobile management terminal is used to maintain a wireless communication connection with the intelligent controller, visualize environmental monitoring data, provide recommendations for suitable plants and identify plant types and growth status, use the Q-learning algorithm to generate personalized maintenance strategies, and build a user content community to incentivize maintenance behavior.

[0010] A soil moisture sensor is installed at the top inside the flowerpot body to monitor the soil moisture inside the flowerpot body in real time.

[0011] A water level sensor is installed on one side of the bottom inside the flowerpot body to monitor the water level data inside the flowerpot body in real time.

[0012] A submersible pump is located on the other side of the bottom inside the flowerpot body, used to draw water stored inside the flowerpot body for plant irrigation;

[0013] The intelligent controller is electrically connected to the soil moisture sensor, water level sensor and submersible pump respectively.

[0014] Furthermore, the flowerpot body includes an outer cover, a tray at the bottom of the outer cover, and top covers on both sides of the top of the outer cover;

[0015] A water tank is installed at the bottom inside the outer cover, and a planting pot is installed at the top inside the outer cover. A partition is installed at the bottom of the planting pot, and a water receiving tray is installed between the top of the water tank and the bottom of the partition. Water spray pipes are installed on both sides of the top of the planting pot.

[0016] A water storage pipe is installed on one side of the partition, between the inner wall of the outer cover and the outer wall of the planting pot. Water is injected into the water tank from top to bottom through the water storage pipe.

[0017] On the other side of the partition, between the inner wall of the water tank and the outer wall of the planting pot, there is a water outlet pipe. The two ends of the water outlet pipe are connected to a submersible pump and a water spray pipe, respectively, which are used to draw water from the bottom of the water tank to the water spray pipe to irrigate the plants inside the planting pot.

[0018] Furthermore, an outer slot is provided on one side of the outer cover, and an inner slot is provided on one side of the top of the water tank. The water receiving tray passes through the outer slot and the inner slot in sequence to the inside of the water tank, and the two side walls of the top of the water tank are provided with sliding rails that cooperate with the water receiving tray.

[0019] The top of the other side of the outer cover has a controller mounting slot that works with the smart controller; the partition is fixedly connected to the bottom of the planting pot, and the bottom of the partition has several drainage holes arranged in a rectangular pattern at equal intervals.

[0020] Furthermore, the intelligent controller includes a microcontroller, a display screen, a clock circuit, a humidity sensor interface, a water level sensor interface, an AC-DC power supply, button input, a memory, a buzzer, a water pump driver, and a communication module.

[0021] The display screen, clock circuit, humidity sensor interface, water level sensor interface, key input, memory, buzzer, water pump driver and communication module are all connected to the microcontroller and AC-DC power supply.

[0022] Furthermore, the mobile management terminal includes:

[0023] The monitoring center module is used to acquire real-time soil moisture and water level data, as well as network-connected light intensity and ambient temperature data, and generate a visual environmental report.

[0024] The knowledge recommendation module is used to identify plant species and growth needs by taking plant images, and to provide disease and pest diagnosis and pesticide recommendations.

[0025] The maintenance execution module is used to analyze plant growth stages and real-time environmental data, dynamically generate personalized maintenance strategies, output control commands, and support users to manually intervene in the execution priority.

[0026] The community interaction module is used to build a user content community, generate maintenance index based on maintenance operations, support users to publish maintenance logs and unlock achievements, and automatically form groups based on user habits;

[0027] The spatial adaptation module is used to build an environment model using augmented reality technology, recommend suitable plants, simulate and preview the future growth state of plants, and support interactive operation and adjustment.

[0028] The monitoring center module is connected to the maintenance execution module and the community interaction module, the maintenance execution module is connected to the knowledge recommendation module and the spatial adaptation module, and the knowledge recommendation module is connected to the spatial adaptation module.

[0029] Furthermore, it analyzes plant growth stages and real-time environmental data to dynamically generate personalized maintenance strategies, output control commands, and support user manual intervention in execution priorities, including:

[0030] The system acquires plant images taken by the user and plant species identification results, extracts the morphological features of the plant, and combines them with the decision tree rule base to determine the current growth stage of the plant. The morphological features include the number of leaves, leaf area index, new leaf germination rate, and flower bud ratio.

[0031] Based on the analysis results of growth stages, a state space for reinforcement learning is constructed with growth stage classification variables as discrete variables and soil moisture gradient values ​​as continuous variables. The hard constraints of the rule engine are activated to ensure that soil moisture is always maintained within the safe humidity range for plants.

[0032] The Q-learning algorithm is used to construct the decision core, the action space is discretized into multiple irrigation amounts and multiple interval durations, the growth index and resource efficiency are integrated to establish a reward function, and a water-saving penalty mechanism is introduced. If the action triggers the rule engine to intercept, a penalty value is applied, and iterative optimization is achieved through continuous interaction.

[0033] The rule engine is used to verify the environmental status. In non-emergency situations, the action combination that maximizes the Q value is selected as the current personalized maintenance strategy for the plant. The control command is generated through format conversion and sent to the smart controller to drive the execution of pulse irrigation.

[0034] Real-time conflict detection is triggered based on user manual operations. By recording the operation content, a user preference adaptive mechanism is formed to update the Q-learning algorithm strategy library.

[0035] Furthermore, based on user manual operations triggering real-time conflict detection, and by recording the operation content, a user preference adaptive mechanism is formed to update the Q-learning algorithm strategy library, including:

[0036] When a user manually triggers a watering operation, conflict detection is immediately initiated, the current automatic personalized maintenance strategy is forcibly interrupted, and the timestamp of the user's operation and the amount of watering are recorded in the database.

[0037] User-manual commands are marked as the highest priority, and automatic watering strategies are set as the middle priority. Timestamp comparisons are used to avoid conflicts with high-weight commands. When resource contention exists, low-weight actions are abandoned.

[0038] User offline operation data is temporarily stored in the database and then uploaded to the cloud, triggering incremental updates to the Q-learning algorithm strategy library.

[0039] Furthermore, a user-generated content community will be established, generating a maintenance index based on maintenance operations. Users will be able to publish maintenance logs and unlock achievements, and automatically form groups based on user habits, including:

[0040] Real-time recording of maintenance operations generates four-dimensional maintenance indicators. A weighted algorithm generates a user-specific maintenance index. Based on a pre-built two-layer community structure, a dual-track driving mechanism combining explicit and implicit incentives is set up to enable user communication and interaction within the user-generated content community.

[0041] The system obtains the types of plant species that users grow, combines them with user behavior characteristics, uses cosine similarity to measure user distance, matches similar plant planting groups, and supports users in publishing and sharing maintenance logs. Among these user behavior characteristics are user operation time preferences, device usage characteristics, and content interaction patterns.

[0042] Furthermore, the two-tiered community structure includes a basic layer and a relational layer. The basic layer provides an interface for publishing maintenance logs in text, image, and video formats, while the relational layer constructs an interest tag tree based on user geographic location and plant species tags.

[0043] The four-dimensional maintenance indicators include the frequency of integrated operations, the diversity of operation types, the improvement value of plant health, and the degree of environmental compatibility.

[0044] Explicit incentives include progress-based, skill-based, and contribution-based incentives;

[0045] The implicit incentive uses contribution value decay, which is calculated by reducing the contribution value based on the user's inactive time, and different levels of activity are divided according to the contribution value.

[0046] Furthermore, augmented reality technology is used to build environmental models, recommend suitable plants, simulate and preview the future growth state of plants, and support interactive adjustments, including:

[0047] Using a mobile terminal to photograph the planting area where the flowerpot is located to achieve environmental perception, calculating the object boundary coordinates through the principle of triangulation, constructing a spatial mapping model, and aligning the virtual coordinate system with the physical space;

[0048] The environmental parameters of the current planting area are matched with the plant growth requirements, and a list of recommended plant species is generated by prioritizing the requirements. The virtual plant model is then loaded into the mobile terminal.

[0049] Load the pre-set flowerpot main model and plant growth keyframe model, use a linear interpolation algorithm to generate a continuous growth animation, map it to the spatial mapping model, and observe whether the plant conflicts with the physical space. If there is a conflict, it is not recommended to plant in the current planting area.

[0050] The beneficial effects of this invention are as follows:

[0051] 1. By adopting an integrated layered design, the system integrates water storage, planting, and irrigation functions. The sliding rail water tray design allows for easy disassembly and cleaning, and the rectangular drainage holes at the bottom of the partition precisely control the amount of water seepage, avoiding the problem of water accumulation and root rot in traditional flower pots. The intelligent controller achieves closed-loop control of environmental data acquisition, irrigation logic judgment, and early warning triggering through fully integrated hardware modules. The AC-DC power supply and memory design ensure that the strategy is not lost after power failure, and the multi-level water level early warning of the buzzer (such as continuous buzzing when the water level is low) significantly reduces the cost of manual monitoring. Thus, through the dual modular design of mechanical and electrical control, the failure rate is effectively reduced, and the maintenance efficiency and portability are improved.

[0052] 2. Based on a dual-track decision-making system of reinforcement learning and rule engine, the system constructs a state space with plant growth stages as discrete variables and soil moisture gradient as a continuous variable. The Q-learning algorithm discretizes irrigation actions into multiple combinations of duration and water volume, and introduces a water-saving penalty mechanism. The rule engine acts as a safety layer to verify the environmental status in real time. When the user intervenes manually, a three-level priority arbitration is immediately triggered (user command > automatic strategy > low-weight action). In case of conflict, inefficient actions are automatically abandoned and operation preferences are recorded to drive the algorithm to incrementally update the strategy library. This effectively improves irrigation accuracy and reduces water waste. Attached Figure Description

[0053] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings:

[0054] Figure 1 This is a system principle block diagram of an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention;

[0055] Figure 2 This is a schematic diagram of the three-dimensional structure of the main body of an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0056] Figure 3 This is a perspective view of the three-dimensional structure of the main body of an automated intelligent flowerpot with water monitoring and early warning functions according to an embodiment of the present invention.

[0057] Figure 4 This is a planar perspective view of the main body of an automated intelligent flowerpot with water monitoring and early warning functions according to an embodiment of the present invention.

[0058] Figure 5 This is a schematic diagram of the outer cover structure of the main body of an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0059] Figure 6This is a schematic diagram of the upper cover structure of the main body of an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0060] Figure 7 This is a schematic diagram of the planting pot structure of an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0061] Figure 8 This is a schematic diagram of the water receiving tray structure of the main body of an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0062] Figure 9 This is a schematic diagram of the water tank structure of the main body of an automated smart flowerpot with water volume monitoring and early warning function according to an embodiment of the present invention.

[0063] Figure 10 This is a schematic diagram of the tray structure of the main body of an automated smart flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0064] Figure 11 This is a schematic diagram of the water spray pipe structure of the main body of an automated intelligent flowerpot with water volume monitoring and early warning function according to an embodiment of the present invention.

[0065] Figure 12 This is a system block diagram of an intelligent controller in an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention;

[0066] Figure 13 This is a circuit diagram of the microcontroller in the intelligent controller of an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0067] Figure 14 This is a circuit diagram of the display screen in the intelligent controller of an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0068] Figure 15 This is a circuit diagram of the clock circuit in the intelligent controller of an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0069] Figure 16 This is a circuit diagram of the humidity sensor interface in an intelligent controller of an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0070] Figure 17 This is a circuit diagram of the water level sensor interface in an intelligent controller of an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0071] Figure 18 This is a circuit diagram of the AC-DC power supply in the intelligent controller of an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0072] Figure 19 This is a circuit diagram of the key input in the intelligent controller of an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0073] Figure 20 This is a circuit diagram of the memory in the intelligent controller of an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0074] Figure 21 This is a circuit diagram of the buzzer in an intelligent controller of an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0075] Figure 22 This is a circuit diagram of the water pump drive in the intelligent controller of an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0076] Figure 23 This is a schematic diagram of a mobile management terminal in an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0077] Figure 24 This is a schematic diagram of the interface of an intelligent controller in an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0078] Figure 25 This is a schematic diagram of the main menu of an automated intelligent flowerpot controller with water monitoring and early warning function according to an embodiment of the present invention;

[0079] Figure 26 This is a schematic diagram of the working mode settings (under the main menu) of an intelligent controller in an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0080] Figure 27 This is a schematic diagram of the humidity quick setting (under the main menu) of the intelligent controller in an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0081] Figure 28 This is a schematic diagram of the settings menu of an intelligent controller in an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0082] Figure 29This is an operational schematic diagram of the control humidity setting of an intelligent controller in an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention;

[0083] Figure 30 This is a schematic diagram illustrating the control period setting operation of an intelligent controller in an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0084] Figure 31 This is a schematic diagram of the interval day cycle of the intelligent controller in an automated intelligent flowerpot with water monitoring and early warning function according to an embodiment of the present invention.

[0085] Figure 32 This is a schematic diagram illustrating the operation of setting the water pump intermittent time in an intelligent controller of an automated intelligent flowerpot with water volume monitoring and early warning function according to an embodiment of the present invention.

[0086] Reference numerals: 1. Flowerpot body; 101. Outer cover; 102. Tray; 103. Top cover; 104. Water tank; 105. Planting pot; 106. Partition; 107. Water tray; 108. Sprayer pipe; 109. Water storage pipe; 110. Water outlet pipe; 111. Outer cavity; 112. Inner cavity; 113. Slide rail; 114. Controller mounting slot; 2. Intelligent controller; 201. Microcontroller; 202. Display screen; 203. Clock circuit; 204. Humidity 205. Sensor interface; 206. Water level sensor interface; 207. AC-DC power supply; 208. Button input; 209. Memory; 210. Buzzer; 211. Water pump driver; 212. Communication module; 3. Mobile management terminal; 301. Monitoring center module; 302. Knowledge recommendation module; 303. Maintenance execution module; 304. Community interaction module; 305. Spatial adaptation module; 4. Soil moisture sensor; 5. Water level sensor; 6. Submersible pump. Detailed Implementation

[0087] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0088] Please see Figures 1-11 An automated smart flowerpot with water level monitoring and early warning functions is provided, comprising:

[0089] The flowerpot body 1 is designed as an integrated unit, combining water storage, planting, and irrigation functions.

[0090] In the description of this invention, the flowerpot body 1 includes an outer cover 101, a tray 102 is provided at the bottom of the outer cover 101, and top covers 103 are provided on both sides of the top of the outer cover 101. A water tank 104 is provided at the bottom inside the outer cover 101, and a planting pot 105 is provided at the top inside the outer cover 101. A partition 106 is provided at the bottom of the planting pot 105. A water receiving tray 107 is provided between the top of the water tank 104 and the bottom of the partition 106. Water spray pipes 108 are provided on both sides of the top of the planting pot 105. The partition 106... A water storage pipe 109 is provided on one side, between the inner wall of the outer cover 101 and the outer wall of the planting pot 105. The water storage pipe 109 fills water into the water tank 104 from top to bottom. On the other side of the partition 106, between the inner wall of the water tank 104 and the outer wall of the planting pot 105, a water outlet pipe 110 is provided. The two ends of the water outlet pipe 110 are connected to a submersible pump 6 and a spray pipe 108, respectively, to draw water from the water tank 104 from bottom to top to the spray pipe 108 to irrigate the plants inside the planting pot 105.

[0091] In the description of this invention, an outer hollow groove 111 is provided on one side of the outer cover 101, and an inner hollow groove 112 is provided on one side of the top of the water tank 104. The water receiving tray 107 passes through the outer hollow groove 111 and the inner hollow groove 112 in sequence to the interior of the water tank 104. Both sides of the top of the water tank 104 are provided with slide rails 113 that cooperate with the water receiving tray 107. The top of the other side of the outer cover 101 is provided with a controller mounting groove 114 that cooperates with the intelligent controller 2. The partition 106 is fixedly connected to the bottom of the planting pot 105, and the bottom of the partition 106 is provided with a plurality of drainage holes arranged in a rectangular pattern at equal intervals.

[0092] It should be noted that the present invention can be customized to a suitable size according to the actual use scenario. For example, for the home balcony planting scenario, an applicable size range can be designed as follows: 400 mm long × 190 mm wide × 160 mm high, with a bottom water tank depth of about 55 mm (capacity 1500 ml). The water storage area and planting area are separated by a 10 mm thick partition. The planting area is about 100 mm deep and can hold about 7 liters of soil. The soil moisture sensor is fixed at the top of the planting area (20 mm from the rim of the pot), and the water level sensor is installed on the side wall of the water tank (10 mm from the bottom). The submersible pump (about 50 mm in diameter) is located on the other side of the water tank and is connected to the spray pipes on both sides through a 6 mm inner diameter water outlet pipe 110. The controller (thickness ≤ 30 mm) is embedded in the mounting groove on the side of the outer cover. The whole is made of lightweight resin material with a thickness of 3 mm, which is suitable for placement on windowsills or balconies.

[0093] The intelligent controller 2 is located on the end face of the flowerpot body 1 and is installed inside the flowerpot body 1. It integrates environmental data acquisition, irrigation logic judgment, early warning triggering and communication interaction functions. It collects soil moisture and water level data, executes control commands and performs negative feedback adjustment and verification, and supports multi-level water volume early warning reminders.

[0094] In the description of this invention, see [reference]. Figures 12-22 The intelligent controller 2 includes a microcontroller 201, a display screen 202, a clock circuit 203, a humidity sensor interface 204, a water level sensor interface 205, an AC-DC power supply 206, a key input 207, a memory 208, a buzzer 209, a water pump driver 210, and a communication module 211. The display screen 202, clock circuit 203, humidity sensor interface 204, water level sensor interface 205, key input 207, memory 208, buzzer 209, water pump driver 210, and communication module 211 are all connected to the microcontroller 201.

[0095] It should be noted that the microcontroller 201 includes: microcontroller chip U1 (model STM32G030C8T6), resistors R1 and R2, capacitors C6 and C7, indicator LED1, and connector J1 (model PIN2.0-4P), and their connection relationship is as follows. Figure 13 As shown; where resistor R2 and capacitor C6 are combined to realize the reset function; resistor R1 and indicator LED1 serve as working indicator lights; connector J1 is used for program download.

[0096] Display screen 202 includes: connector JU2 (model PIN2.54-8P) and OLED display module U2 (model ZJY130S08Z0WG01), and their connection relationship is as follows: Figure 14 As shown.

[0097] Clock circuit 203 includes: a battery, a battery holder, capacitors C1, C2, and C16, and a crystal oscillator X1, whose connection relationship is as follows: Figure 15 As shown.

[0098] The humidity sensor interface 204 includes: audio jack PH1, thermistor PTC1, resistors R3, R8, R14, R15, and capacitor C8; it is a serial port sensor by default. R15 and R14 should be soldered as standard (No BOM). The connection relationship is as follows: Figure 16 As shown.

[0099] The water level sensor interface 205 includes: an audio jack PH2, resistors R6, R7, and R13, and capacitors C17 and C18; their connection relationships are as follows: Figure 17 As shown, when the water shortage sensor is connected, the W_HIGH port is set to push-pull output to power the sensor.

[0100] AC-DC power supply 206 includes: power module MU3 (model ACDC-JZ-3W), connector J4, inductor L1, capacitors C11, C12, C13, C14, and C15, bidirectional diode TVS1 (model SMAJ6.8CA), linear regulator U4 (model SPX1117-3.3), varistor MOV1 (model 07D471K), and fuse FU1; their connection relationships are as follows. Figure 18 As shown.

[0101] Key input 207 includes: connector J2, resistors R9, R10, R11, and R12, capacitors C19, C20, C21, and C22; their connection relationship is as follows. Figure 19 As shown.

[0102] Memory 208 includes: memory U3 (model AT24C04XC-STUM-T) and capacitor C3; their connection relationship is as follows: Figure 20 As shown.

[0103] Buzzer 209 includes: buzzer BZ1, diode D2, transistor Q2, resistor R8, capacitor C4, and capacitor C5; their connection relationship is as follows. Figure 21 As shown.

[0104] The water pump drive 210 includes: relay RLY1B (model HF49FD005-1H11), relay RLY1A (model HF49FD005-1H11), indicator LED2, capacitor C9, capacitor C10, resistor R4, resistor R5, and transistor Q1; their connection relationship is as follows. Figure 22 As shown.

[0105] In the description of this invention, as Figure 24 As shown in the diagram, the interface of the intelligent controller 2 is illustrated. The intelligent controller 2 automatically controls the water pump to turn on or off based on the measured soil moisture value, so that the soil moisture is always kept within the set humidity range.

[0106] The water level sensor interface of the intelligent controller 2 can be connected to the water level sensor 5. Its main purpose is to prevent the water level in the tank from falling below the water pump, which could cause the pump to run dry and burn out. When the water level sensor 5 detects that the water level in the tank is lower than the water pump, the water pump is prevented from starting. Both water level interfaces function identically; the water level sensor 5 can be connected to the interface closer to the water tank.

[0107] The main functions of the intelligent controller 2 include:

[0108] 1. Soil moisture testing: Detection of substrate moisture content;

[0109] 2. Humidity control mode: Automatically waters the substrate based on the set moisture content;

[0110] 3. Timed watering control mode: Automatically waters according to the set time period and interval;

[0111] 4. Timed + Humidity Control Mode: Watering begins during the set time period when the substrate moisture content is below the set value;

[0112] 5. Intermittent control mode: The water pump is controlled cyclically according to the set opening and closing times, and can be used in conjunction with other modes;

[0113] 6. Manual control mode: Manually control the water pump to turn on and off;

[0114] 7. Water pump water shortage protection: The water pump will automatically shut down and an alarm will be sound when the water tank is low on water (water shortage sensor is an optional accessory).

[0115] 8. Do Not Disturb Alarm: The alarm sound will automatically turn off at night.

[0116] The operating parameters of the intelligent controller 2 include:

[0117] 1. Power input: AC 220V;

[0118] 2. Water pump output power: AC 220V (the water pump power must be less than 300W, and high-power electrical appliances cannot be connected to it).

[0119] 3. Soil moisture measurement range: 0%~100%;

[0120] 4. Soil moisture setting range: 1%~99%;

[0121] 5. Timed control range: Two timed segments throughout the day, with an interval of one day or 1 to 99 days;

[0122] 6. Intermittent control time: 1 to 999 seconds.

[0123] The intelligent controller 2 has four operating modes: manual control mode, timer control mode, humidity control mode, and timer + humidity control mode. The control principles of each mode are as follows:

[0124] 1. Manual control mode: The water pump can be directly controlled via the water pump switch.

[0125] 2. Timed control mode: It can automatically water according to the set time and interval days. Up to 2 timed times can be set per day. When the time reaches the set time range, the controller turns on the water pump to start watering. When the time exceeds the set time range, the controller turns off the water pump to stop watering.

[0126] 3. Humidity control mode: It can automatically water according to the set soil moisture, with a soil moisture setting range of 0% to 99%; when the soil moisture is less than the minimum set value, the controller turns on the water pump to start watering; when the soil moisture is greater than the maximum set value, the controller turns off the water pump to stop watering.

[0127] 4. Timed + Humidity Control Mode: Automatic watering based on set time and soil moisture; when the timer reaches the set time range and the soil moisture is less than the set value, the controller turns on the water pump to start watering; when the timer exceeds the set time range and the soil moisture is greater than the set value, the controller turns off the water pump to stop watering.

[0128] In addition, the intelligent controller 2 also supports intermittent control: it can cycle the water pump on or off according to the set on and off durations. When the intermittent control function is activated, the water pump will operate in an intermittent control manner during the on period in all operating modes.

[0129] It should be noted that in manual control mode, the submersible pump can be directly operated via the switch button, or the user can click the pump switch button on the visual interface. The command is transmitted to the smart controller via Wi-Fi / Bluetooth, triggering the highest priority interrupt mechanism. At this time, the automatic strategy of the maintenance execution module 303 is forcibly suspended, and the smart controller 2 only performs immediate switching actions. Meanwhile, the mobile terminal displays the pump status changes in real time and records the operation timestamp, and updates the Q-learning algorithm strategy library through the user preference adaptive mechanism.

[0130] The timed control mode is configured through the maintenance execution module 303 of the mobile management terminal 3 or the operation button of the smart controller 2. That is, the user sets the irrigation period (such as "7:00-7:30 every day") and repetition cycle through the maintenance execution module 303. The parameters are synchronized by the clock circuit to generate a timed task queue. The mobile terminal encapsulates the time period instruction into JSON format and stores it in the SQLite database. When the network fluctuates, the local cache is automatically activated. When the preset time is reached, the smart controller 2 accurately triggers the submersible pump 6 to drive. At the same time, the community interaction module 304 records the user's operating habits to optimize future time period planning.

[0131] The humidity control mode is supported by the knowledge recommendation module 302 of the mobile management terminal 3, which provides dynamic threshold support. It calls the plant database to match the water requirement characteristics of the species (such as setting the humidity threshold of succulents to 30%~50%), and generates the irrigation threshold range by combining the real-time soil capacitance and humidity value. When the soil moisture sensor data is lower than the set lower limit, the controller starts the submersible pump 6 and stops the pump when it is higher than the upper limit. The mobile terminal simultaneously marks the abnormal parameters and pushes the alarm to the visualization interface. At the same time, the maintenance execution module 303 continuously optimizes the threshold accuracy through reinforcement learning algorithm.

[0132] The timed + humidity control mode requires multi-module collaborative decision-making of the mobile management terminal 3. The maintenance execution module 303 binds the time interval and humidity threshold to generate composite instructions (such as "start only on weekday mornings and humidity <40%)". The community interaction module 304 analyzes the user's historical operations to improve the strategy matching degree. The intelligent controller 2 executes a dual-condition arbitration mechanism - irrigation is triggered only when the timed conditions are met and the humidity is lower than the threshold. If either condition fails, it enters sleep mode. In case of an anomaly, the mobile terminal automatically switches to single mode and pushes an adjustment plan.

[0133] The main menu and some standalone operations of the intelligent controller 2 are as follows: Figures 25-32 As shown.

[0134] like Figure 26 As shown, for the working mode setting (under the main menu), pressing the "On | Mode" button in the main menu can switch the controller's working mode; in any working mode, pressing the "Pump Off" button can exit the mode and turn off the water pump.

[0135] like Figure 27 As shown, for quick humidity setting (under the main menu), you can press the "Set Humidity (Confirm)" button to directly set the humidity value without entering the settings menu; press the "Confirm" button to switch between the humidity when the pump is on and the humidity when the pump is off; press "+" and "-" to adjust the humidity value; press the "Exit" button to exit the settings.

[0136] like Figure 28 As shown, the system settings menu allows you to set the humidity control value, control period, water pump intermittent control time, and the controller's time and date. Press and hold the "Settings" button in the main menu to enter the system settings menu; press the "Settings" button again to return to the main menu.

[0137] like Figure 29 The image shows the humidity control settings. Humidity can also be set directly in the main menu.

[0138] like Figure 30 The image shows the control time period settings.

[0139] like Figure 31 As shown, this is a cyclical pattern for the interval number of days.

[0140] like Figure 32 As shown, this is the setting for the water pump intermittent time.

[0141] Mobile management terminal 3 is used to maintain wireless communication with intelligent controller 2, visualize environmental monitoring data, provide recommendations for suitable plants and identification of plant types and growth status, generate personalized maintenance strategies using Q-learning algorithm, and build a user content community to incentivize maintenance behavior.

[0142] In the description of this invention, as Figure 23As shown, the mobile management terminal 3 includes:

[0143] The monitoring center module 301 is used to acquire real-time soil moisture and water level data, as well as network-connected light intensity and ambient temperature, and generate a visual environmental report.

[0144] Specifically, the mobile terminal connects to the built-in smart controller 2 of the smart flowerpot or the corresponding soil moisture sensor and water level sensor via Bluetooth or Wi-Fi protocol to obtain data in real time; at the same time, it calls the GPS positioning and ambient light sensor of the mobile terminal, or connects to third-party meteorological application software to supplement light intensity and ambient temperature data.

[0145] It uses a time-series database to store dynamic data streams and combines a data visualization library to generate dynamic chart reports, including real-time data curves, environmental parameter heatmaps, and historical trend comparisons. Users can export PDF reports with one click.

[0146] The knowledge recommendation module 302 is used to identify plant species and growth needs by taking plant images, and to provide disease and pest diagnosis and pesticide recommendations.

[0147] Specifically, a lightweight SQLite database is deployed on mobile devices to cache frequently accessed data on common houseplants (including basic growth parameters and pest and disease characteristic maps); a distributed MongoDB database is used in the cloud to store the full range of species data, and bidirectional data updates are achieved through an incremental synchronization protocol. In offline mode, the system automatically switches to the local database to ensure the availability of basic identification functions.

[0148] To identify plant species and growth requirements through plant images, and to provide disease and pest diagnosis and pesticide recommendations, an image processing scheme based on convolutional neural networks can be used, including the following:

[0149] 1. Users use the camera of their mobile terminal to acquire high-definition images and use image preprocessing techniques (such as histogram equalization to adjust contrast and random erasure to enhance robustness) to optimize image quality.

[0150] 2. Use a convolutional neural network (such as MobileNetV3 or YOLOv8) to segment the main area of ​​the plant, extract morphological features (such as leaf vein texture, flower bud shape, and lesion distribution) and generate feature vectors.

[0151] 3. The feature vectors are matched with a pre-built plant species database, and color histograms and texture features are integrated for similarity matching. The species and growth stage are determined by combining the decision tree rule base. For example, the leaf area index (LAI) is used to distinguish between the seedling stage and the flowering stage.

[0152] 4. In the disease and pest diagnosis stage, target detection is performed based on the segmented lesion areas to locate abnormal areas. The disease type is identified by comparing with the disease and pest image database. At the same time, the correlation between the pathogenic factors and the environment is analyzed using the knowledge graph.

[0153] 5. Based on the diagnostic results, a safe and effective prevention and control plan is matched through a knowledge base such as a drug database or a large language model. The dosage and application frequency of the drug are dynamically adjusted by integrating real-time environmental data, and natural language guidance text is generated and fed back to the user interface.

[0154] During operation, multi-scale feature extraction (leaf texture / flower shape, etc.) is performed on plant images taken by users. Species identification is achieved by comparing with the database. Then, the lesion spread morphology is analyzed by combining the HSV color space and the pathogen type is output by associating with the database.

[0155] The maintenance execution module 303 is used to analyze plant growth stages and real-time environmental data, dynamically generate personalized maintenance strategies, output control commands, and support users to manually intervene in the execution priority.

[0156] In the description of this invention, the plant growth stage and real-time environmental data are analyzed to dynamically generate personalized maintenance strategies, output control commands, and support user manual intervention. The execution priority includes:

[0157] Step S11: Obtain the plant image taken by the user and the plant species identification result, extract the morphological features of the plant, and determine the current growth stage of the plant by combining the decision tree rule base. The morphological features include the number of leaves, leaf area index (LAI), new leaf germination rate and flower bud ratio.

[0158] Specifically, the growth stage is analyzed through a pre-set decision tree rule base. For example, when the LAI value is below 0.5, it is determined to be in the seedling stage because the root system has weak water absorption capacity at this stage and a small amount of water is required for high-frequency irrigation. When the LAI value is between 0.5 and 2.0, it is marked as the growth stage because the water requirement increases with the growth of biomass. When the LAI is ≥ 2.0 and the proportion of flower buds exceeds 5%, it is classified as the flowering stage because water is sensitive at this stage and water shortage must be prevented to avoid bud drop.

[0159] Step S12: Based on the analysis results of the growth stage, construct the state space of reinforcement learning with the growth stage classification variable as the discrete variable and the soil moisture gradient value as the continuous variable (the moisture gradient is calculated from the spatial standard deviation of the soil moisture in the flower pot; if it is greater than 15%, it is marked as uneven water absorption by the roots). Then, activate the hard constraints of the rule engine to ensure that the soil moisture is always maintained within the safe moisture range for plants.

[0160] Specifically, the hard constraint mechanism of the rule engine is activated, such as defining safety rules to force the irrigation action when the soil moisture is below 30% (to avoid plant dehydration and death), embedding crop rules to limit the single irrigation amount during the flowering period to ≤150ml (to prevent water shock from causing bud drop) and the irrigation interval during the seedling period to ≥4 hours (to avoid the risk of root rot), and using the rule engine to scan the action space in real time and intercept action parameters that violate the constraints (such as automatically downgrading to the compliant value if an irrigation instruction of >150ml is generated during the flowering period).

[0161] Step S13: The Q-learning algorithm is used to construct the decision core. The action space is discretized into multiple irrigation amounts and multiple interval durations. The growth index and resource efficiency are integrated to establish a reward function. A water-saving penalty mechanism is introduced. If the action triggers the rule engine to intercept, a penalty value is applied. Iterative optimization is achieved through continuous interaction.

[0162] Specifically, the Q-learning algorithm serves as the core decision-making mechanism in the personalized maintenance strategy of smart flowerpots. It optimizes irrigation action selection by simulating the continuous interaction between the agent and the plant's growth environment. The algorithm integrates plant growth stages (such as seedling stage, growth stage, and flowering stage) and environmental parameters such as soil moisture gradient into a discrete state space, and defines irrigation actions (such as different watering amounts and intervals) as a discrete action space. It learns the optimal strategy by iteratively updating the long-term value prediction (Q-value) of state-action pairs.

[0163] In practical applications of this invention, the Q-learning algorithm integrates plant growth indicators (such as leaf development and bud status) with resource efficiency to construct a multi-dimensional reward function. For example, when the system detects an improvement in plant health, it provides positive incentives. At the same time, it introduces water-saving penalties and rule-based interception penalties to avoid ineffective or high-risk actions, ensuring that the strategy meets both the physiological needs of plants and takes into account sustainability.

[0164] Its functions are mainly reflected in three aspects:

[0165] 1. Dynamic strategy generation: Based on the real-time analysis of the growth stage and environmental conditions, select the irrigation combination that maximizes long-term rewards from the action space to replace the fixed rule strategy. For example, automatically reduce the amount of water supplied at one time during the flowering period to avoid the risk of bud drop, and increase the irrigation frequency during the growth period to support biomass accumulation.

[0166] 2. Safety Constraint Embedding: Working in conjunction with the rule engine, hard constraints (such as soil moisture safety thresholds) directly limit the range of action choices, while the algorithm learns to actively avoid violations through a penalty mechanism, forming a double guarantee;

[0167] 3. User preference adaptation: When the user intervenes manually, the recorded operation data triggers an incremental update of the Q value, so that subsequent strategies gradually align with the user's habits. For example, if the user frequently drinks water during a specific period, the algorithm will adjust the status action value of that period to increase its priority.

[0168] The core effects include: continuously learning the water requirements of different plant varieties and growth stages through interactive learning, dynamically optimizing irrigation parameters, and reducing the failure rate of maintenance caused by sudden environmental changes or variety differences; the water-saving penalty mechanism in the reward function drives the algorithm to minimize water consumption while ensuring plant health, for example by exploring low-water-consumption strategies by extending irrigation intervals; and combined with real-time verification by the rule engine, it avoids triggering harmful actions (such as excessive irrigation during flowering) during the algorithm's exploration process, while accelerating the elimination of inefficient strategies through the penalty mechanism.

[0169] For example, the action space is discretized into different levels of irrigation volume (e.g., 20ml / 50ml / 80ml / 110ml / 140ml) and corresponding interval durations (e.g., 2h / 4h / 6h / 8h / 12h). The reward function design integrates growth indicators and resource efficiency: a reward of +1 is given for every 0.1 increase in LAI during the seedling stage, and a reward of +2 is given for every 5% increase in flower bud opening rate during the flowering stage; at the same time, a water-saving penalty mechanism is introduced (reward points are deducted according to the excess proportion when the water consumption per unit of organism exceeds the threshold, the formula is -0.5 × excess percentage), and a penalty value of -5 is applied if the action triggers the rule engine to intercept it. The strategy is iteratively optimized through continuous interaction and updating of the Q-table.

[0170] Step S14: Use the rule engine to verify the environmental status, select the action combination that maximizes the Q value in non-emergency state as the current personalized maintenance strategy for the plant, and generate control instructions through format conversion, and send them to the smart controller 2 to drive the execution of pulse irrigation.

[0171] Step S15: Real-time conflict detection is triggered based on user manual operation. By recording the operation content, a user preference adaptive mechanism is formed, and the Q-learning algorithm strategy library is updated.

[0172] In the description of this invention, real-time conflict detection is triggered based on user manual operations. By recording the operation content, a user preference adaptive mechanism is formed, and the Q-learning algorithm strategy library is updated, including:

[0173] Step S151: When the user manually triggers the watering operation, conflict detection is immediately started, the current automatic personalized maintenance strategy is forcibly interrupted, and the timestamp of the user's operation and the amount of irrigation are recorded in the database.

[0174] Step S152: Mark the user's manual command as the highest priority and the automatic watering strategy as the middle priority. Avoid conflicts with high-weight commands by comparing timestamps. When there is resource competition, abandon low-weight actions.

[0175] Step S153: Temporarily store the user's offline operation data in the database, then upload it to the cloud to trigger an incremental update of the Q-learning algorithm strategy library (Q-table).

[0176] The Q-learning algorithm strategy library is the decision matrix of the Q-learning algorithm, with the following structure: Rows: 30 discretized states (such as combinations of plant growth stages and soil moisture levels); Columns: 3 watering actions (move left / stay / move right); Cell value (Q value): representing the long-term profit prediction of performing a certain action in a specific state, which is dynamically updated through a reward function.

[0177] The community interaction module 304 is used to establish a user content community, generate a maintenance index based on maintenance operations, support users to publish maintenance logs and unlock achievements, and automatically form groups based on user habits.

[0178] In the description of this invention, a user content community is established, a maintenance index is generated based on maintenance operations, users are supported in publishing maintenance logs and unlocking achievements, and automatic grouping based on user habits includes:

[0179] Step S21: Record maintenance operation records in real time, generate four-dimensional maintenance indicators (including the frequency of integrated operations, the diversity of operation types, the improvement value of plant health and the degree of environmental matching), generate a user-specific maintenance index through a weighted algorithm, and build a two-layer community structure. Set up a dual-track driving mechanism that combines explicit incentives and implicit incentives to realize user communication and interaction within the user content community.

[0180] Specifically, the operation frequency dimension refers to the real-time recording of timestamps and frequency of user operations such as watering, fertilizing, and pruning via IoT sensors, calculating the operation frequency per unit time. The operation variety dimension refers to the statistical coverage of operation categories performed by the user based on a preset maintenance operation classification system. For example, a user simultaneously performing watering, repotting, and supplemental lighting will have a higher diversity score than a user performing only watering. The plant health improvement value dimension refers to the analysis of plant morphological changes through image recognition models, including leaf number growth rate, pest and disease elimination rate, and chlorophyll index improvement. The environmental matching degree dimension refers to comparing the differences between the actual environment and the plant's needs model.

[0181] In the process of calculating the maintenance index, each of the above four maintenance indicators is assigned its own weight, and the user's maintenance index is obtained by weighted summation. The weight baseline is operation frequency 25%, diversity 15%, health 40%, and environment 20%.

[0182] In the description of this invention, the two-layer community structure includes: a basic layer that provides an open interface for publishing maintenance logs in text / image / video formats, and uses a distributed object storage system to manage media files to ensure content accessibility in high-concurrency scenarios; and a relationship layer that constructs an interest tag tree based on user geographic location and plant type tags to provide a classification basis for subsequent automatic grouping.

[0183] The two-tier community structure can be built on commonly used development platforms, such as the WeChat Mini Program ecosystem, other application Mini Programs, or open platforms.

[0184] In building a two-tiered community within the WeChat Mini Program ecosystem, the foundational layer leverages the open capabilities of Mini Programs to implement interfaces for publishing text / image / video logs. Media files are managed through cloud object storage, and CDN (Content Delivery Network, a technology that improves content delivery efficiency through a distributed network system) is used to accelerate high-concurrency access performance. User-uploaded maintenance logs are compressed and stored on distributed storage nodes, linked to user accounts via unique file IDs, supporting tens of thousands of concurrent reads. The relationship layer requires calling location application interfaces to obtain the user's authorized geographic coordinates, combining this with user-manually labeled plant type tags to construct a dynamic interest tag tree. For example, user A and user B are automatically matched to "a certain region - shade-tolerant plant community".

[0185] Explicit incentives pre-set three types of achievements: progress-based, skill-based, and contribution-based, with achievement unlocking conditions tied to maintenance index thresholds. Implicit incentives introduce a contribution value decay algorithm, where contribution values ​​decrease exponentially with user inactivity. For example, users ranking in the top 10% by contribution value are automatically granted "community mentor" status and given priority access to content recommendations.

[0186] It should be noted that explicit incentives in the software rely on a rule engine and real-time data stream processing. Specifically, progress-based achievements (such as continuous maintenance check-ins) record the frequency of user behavior through a Redis key-value accumulator, and an achievement unlock event is triggered when 7 consecutive check-ins are completed. Skill-based achievements (such as saving dying plants) need to be integrated with environmental sensor APIs (temperature, humidity / light intensity). When the maintenance index is ≥80 points and the rescue action is completed 3 times, the achievement is activated after real-time calculation and verification. Contribution-based achievements (such as a solution being adopted) rely on a community voting system. The adoption count is stored in the document database, and the achievement is granted after 10 adoptions are reached through a query. All achievement data is stored in a cloud database, and the achievement status is synchronized to the 3D rendering engine in real time, dynamically generating 3D badges on the user's AR interface.

[0187] The core of implicit incentives lies in the dynamic contribution value system and decay algorithm. The initial calculation of user contribution values ​​is based on explicit achievement weights (e.g., skill-based achievement contribution value +20), supporting real-time ranking. Inactive users (no login / posting behavior) are processed in batches daily. When there is no interaction for 30 consecutive days, λ=0.05 (daily decay of 5%), and the contribution values ​​are automatically re-ranked after updates. For example, the top 10% of users can be awarded the "Community Mentor" identity, which is associated with a content recommendation weight coefficient (1.0 for ordinary content and 1.5 for mentor content), allowing mentor content to be displayed preferentially within the community. If a user falls out of the top 10% due to decay, their interest tags (e.g., "succulent care") are retrieved from the graph database, and unresolved help tasks in related communities are automatically pushed to them. Completing the task resets the decay timer.

[0188] Step S22: Obtain the plant species categories (flowering / foliage / succulent distribution) planted by the user, combine user behavior characteristics, use cosine similarity to measure user distance, match similar plant planting groups, and support users to publish and share maintenance logs.

[0189] Among them, user behavior characteristics include operation time preferences (morning / nighttime activity), device usage characteristics (mobile / desktop ratio), and content interaction patterns (like / comment / share ratio).

[0190] The spatial adaptation module 305 is used to build an environmental model using augmented reality technology, recommend suitable plants, simulate and preview the future growth state of plants, and support interactive operation and adjustment.

[0191] In the description of this invention, an environmental model is established using augmented reality technology to recommend suitable plants, simulate and preview the future growth state of the plants, and support interactive operation adjustments, including:

[0192] Step S31: Use a mobile terminal to capture images of the planting area for environmental perception, calculate the boundary coordinates of objects using the principle of triangulation, construct a simple spatial mapping model, and align the virtual coordinate system with the physical space. Specifically, the mobile terminal's camera captures real-time images of the planting area where the flowerpot is located, and combines this with built-in sensors such as accelerometers and gyroscopes to perform multi-source data fusion; the relative distances and boundary coordinates of objects in the environment are calculated using the principle of triangulation.

[0193] It should be noted that for environmental modeling and constructing spatial mapping models, the SLAM algorithm (Simultaneous Localization and Mapping) can be used. This algorithm captures environmental images through the mobile terminal's camera and combines them with sensor data from gyroscopes, accelerometers, and other sensors to calculate the spatial position of objects in real time. Augmented reality software, such as ARKi (Augmented Reality Development Kit) or ARCore (Augmented Reality Development Platform), provides the core framework. This functional interface is then connected to the open interface of the already constructed two-layer community structure. Its planar detection function is used to identify boundaries such as the ground and walls, generating 3D point cloud data.

[0194] For applications combining 3D, such as the Unity engine (a cross-platform game engine), this engine integrates point cloud data and converts it into a visual mesh model, supporting the alignment of the virtual coordinate system with physical space, and rendering a spatial framework that includes obstacle positions and lighting gradients.

[0195] Step S32: Match the environmental parameters of the current planting area with the plant growth requirements, generate a recommended list of plant species by prioritizing the requirements, and load the virtual plant model in the augmented reality interface of the mobile terminal.

[0196] Specifically, environmental parameters extracted from the spatial model (such as light intensity, temperature and humidity range, and physical space size) are matched with a pre-set database of plant growth requirements; a multi-objective decision-making algorithm is used to quantify and score plant adaptability, and a ranking list is generated with the priority of requirements (such as shade tolerance, spatial adaptability, and maintenance difficulty) as weights.

[0197] Virtual plant models are loaded into a spatial mapping model constructed through a mobile terminal. The model is dynamically generated based on the morphological characteristics of the plant species (such as plant height, crown width, and root morphology) and superimposed onto the corresponding coordinate position of the spatial mapping model. Users can rotate or zoom the virtual model through gesture interaction to observe its spatial adaptability to the environment.

[0198] Step S33: Load the preset flowerpot body 1 model and plant growth keyframe model, use linear interpolation algorithm to generate continuous growth animation, map it to the spatial mapping model, observe whether the plant conflicts with the physical space, if there is a conflict, it is not recommended to plant in the current planting area; and support gesture operation, drag the virtual plant model to the transplantable area.

[0199] Specifically, by calling pre-stored keyframe data of plant growth (such as 3D models of the budding, growth, and maturity stages), a linear interpolation algorithm is used to generate a smooth, continuous growth animation between keyframes, simulating the morphological changes of the plant from its current state to several months in the future. This animation sequence is then mapped in real time to a spatial mapping model, and a collision detection algorithm is used to analyze the geometric intersection of the virtual plant's branches and leaves with obstacles in the environment (such as walls and furniture) frame by frame. If branches and leaves are detected penetrating obstacles or the light occlusion rate exceeds a threshold, it is determined to be a spatial conflict, and the species is automatically removed from the recommendation list. The conflict area is highlighted with color, allowing users to manually adjust the position of the flowerpot or the environmental layout to re-verify the suitability.

[0200] Furthermore, the plant growth process simulation is applicable to the analysis of the conventional growth trajectories of various common plant types, but it will not perfectly match the characteristics of the currently planted plants and should only be used as a reference function. Linear interpolation algorithms have been widely validated in the field of numerical computation for their high computational efficiency, simple implementation, and suitability for real-time rendering on mobile devices. Although it cannot completely simulate the characteristics of individual plants, it can provide a reliable morphological trend reference by covering conventional growth trajectories with keyframe data, and conflict detection can meet the spatial safety prediction needs of home scenarios. Similar technologies have been applied to AR projects such as "Tracing the Origins of Herbs," where flashlight illumination triggers plant growth animations and real-time environmental adaptability detection verifies the feasibility of the function's implementation.

[0201] It should be noted that user operation records using augmented reality technology (such as spatial models and conflict points) are stored in a distributed object system to generate reusable planting plan templates. These templates are linked to a community log system, supporting sharing in text / image / video formats, and accelerated for high-concurrency access via a distributed network. Based on user environment tags and maintenance indices, interest-based communities are built, and mature plant augmented reality models from users in the same city are automatically pushed to them.

[0202] Soil moisture sensor 4 (TEROS 11, MSE or SMT100 can be selected) is used to monitor the soil moisture inside the flowerpot body 1 in real time.

[0203] Water level sensor 5 (such as PY201 submersible liquid level sensor or DP-FYC-3 float water level sensor) is used to monitor the water level data inside the flowerpot body 1 in real time.

[0204] Submersible pump 6 (QDX series miniature submersible pump, QJR stainless steel submersible pump, etc. can be selected) is used to draw water stored inside the flowerpot body 1 for plant irrigation.

[0205] In summary, by employing the above-mentioned technical solution of this invention, the integrated layered design achieves the integration of water storage, planting, and irrigation functions. The sliding rail water tray design supports easy disassembly and cleaning, and the rectangular drainage holes at the bottom of the partition precisely control the amount of water seepage, avoiding the problem of water accumulation and root rot in traditional flower pots. The intelligent controller achieves closed-loop control of environmental data acquisition, irrigation logic judgment, and early warning triggering through the full integration of hardware modules. The AC-DC power supply and memory design ensure that the strategy is not lost after power failure, and the multi-level water level early warning of the buzzer (such as continuous buzzing at low water level) significantly reduces the cost of manual monitoring. Thus, through the dual modular design of mechanical and electrical control, the failure rate is effectively reduced, and maintenance efficiency and portability are improved. Based on a dual-track decision-making approach combining reinforcement learning and a rule engine, a state space is constructed using plant growth stages as discrete variables and soil moisture gradient as a continuous variable. The Q-learning algorithm discretizes irrigation actions into multiple combinations of duration and water volume, and introduces a water-saving penalty mechanism. The rule engine acts as a safety layer to verify the environmental state in real time. When the user intervenes manually, a three-level priority arbitration is immediately triggered (user command > automatic strategy > low-weight action). In case of conflict, inefficient actions are automatically abandoned and operation preferences are recorded to drive the algorithm to incrementally update the strategy library. This effectively improves irrigation accuracy and reduces water waste.

[0206] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

Claims

1. An automated intelligent flowerpot with water volume monitoring and early warning function, characterized in that, include: The main body of the flowerpot is designed in an integrated manner, combining water storage, planting and irrigation functions; The intelligent controller is set on the end face of the flowerpot body and is used to integrate environmental data acquisition, irrigation logic judgment, early warning triggering and communication interaction functions. It collects soil moisture and water level data, executes control commands and performs negative feedback adjustment and verification, and supports multi-level water volume early warning reminders. A mobile management terminal is used to maintain a wireless communication connection with the intelligent controller, visualize environmental monitoring data, provide recommendations for suitable plants and identify plant types and growth status, generate personalized maintenance strategies using the Q-learning algorithm, and build a user content community to incentivize maintenance behavior; a soil moisture sensor is located at the top inside the flowerpot body to monitor the soil moisture inside the flowerpot body in real time; a water level sensor is located on one side of the bottom inside the flowerpot body to monitor the water level data inside the flowerpot body in real time; a submersible pump is located on the other side of the bottom inside the flowerpot body to draw water stored inside the flowerpot body for plant irrigation; the intelligent controller maintains electrical connections with the soil moisture sensor, water level sensor, and submersible pump respectively. The mobile management terminal includes: a maintenance execution module, used to analyze plant growth stages and real-time environmental data, dynamically generate personalized maintenance strategies, output control commands, and support user manual intervention in execution priority; a community interaction module, used to establish a user content community, generate maintenance indices based on maintenance operations, support users to publish maintenance logs and unlock achievements, and automatically form groups based on user habits; and a spatial adaptation module, used to use augmented reality technology to build environmental models, recommend suitable plants, simulate and preview the future growth state of plants, and support interactive adjustments; and analyze plant growth stages and real-time environmental data to dynamically generate personalized maintenance strategies, output control commands, and support user manual intervention in execution priority. The system includes: constructing a decision core using the Q-learning algorithm, discretizing the action space into multiple irrigation volumes and interval durations, establishing a reward function by integrating growth indicators and resource efficiency, and introducing a water-saving penalty mechanism. If an action triggers an interception by the rule engine, a penalty value is applied, and iterative optimization is achieved through continuous interaction; the rule engine is used to verify the environmental status, selecting the action combination that maximizes the Q value in non-emergency states as the current personalized maintenance strategy for the plants, and generating control instructions through format conversion, which are then sent to the intelligent controller to drive pulsed irrigation; real-time conflict detection is triggered based on user manual operations, and by recording the operation content, a user preference adaptive mechanism is formed to update the Q-learning algorithm strategy library. Based on real-time conflict detection triggered by user manual operations, and by recording the operation content, a user preference adaptive mechanism is formed to update the Q-learning algorithm strategy library, including: when a user manually triggers a watering operation, conflict detection is immediately initiated, forcibly interrupting the current automatic personalized maintenance strategy, and recording the timestamp and irrigation amount of the user operation in the database; the user's manual command is marked as the highest priority, and the automatic watering strategy is marked as the middle priority, avoiding conflicts with high-weight commands by comparing timestamps, and abandoning low-weight actions when resource competition exists; the user's offline operation data is temporarily stored in the database and then uploaded to the cloud, triggering incremental updates of the Q-learning algorithm strategy library.

2. The automated intelligent flowerpot with water monitoring and early warning function according to claim 1, characterized in that, The flowerpot body includes an outer cover, a tray at the bottom of the outer cover, and top covers on both sides of the top of the outer cover; A water tank is installed at the bottom inside the outer cover, and a planting pot is installed at the top inside the outer cover. A partition is installed at the bottom of the planting pot, and a water receiving tray is installed between the top of the water tank and the bottom of the partition. Water spray pipes are installed on both sides of the top of the planting pot. A water storage pipe is installed on one side of the partition, between the inner wall of the outer cover and the outer wall of the planting pot. Water is injected into the water tank from top to bottom through the water storage pipe. On the other side of the partition, between the inner wall of the water tank and the outer wall of the planting pot, there is a water outlet pipe. The two ends of the water outlet pipe are connected to a submersible pump and a water spray pipe, respectively, which are used to draw water from the bottom of the water tank to the water spray pipe to irrigate the plants inside the planting pot.

3. An automated intelligent flowerpot with water monitoring and early warning function according to claim 2, characterized in that, An outer slot is provided on one side of the outer cover, and an inner slot is provided on one side of the top of the water tank. The water receiving tray passes through the outer slot and the inner slot in sequence to the inside of the water tank. Slide rails that cooperate with the water receiving tray are provided on both sides of the top of the water tank. The top of the other side of the outer cover has a controller mounting slot that works with the smart controller; the partition is fixedly connected to the bottom of the planting pot, and the bottom of the partition has several drainage holes arranged in a rectangular pattern at equal intervals.

4. An automated intelligent flowerpot with water monitoring and early warning function according to claim 1, characterized in that, The intelligent controller includes a microcontroller, display screen, clock circuit, humidity sensor interface, water level sensor interface, AC-DC power supply, key input, memory, buzzer, water pump drive and communication module; The display screen, clock circuit, humidity sensor interface, water level sensor interface, key input, memory, buzzer, water pump driver and communication module are all connected to the microcontroller and AC-DC power supply.

5. An automated intelligent flowerpot with water monitoring and early warning function according to claim 1, characterized in that, The mobile management terminal also includes: The monitoring center module is used to acquire real-time soil moisture and water level data, as well as network-connected light intensity and ambient temperature data, and generate a visual environmental report. The knowledge recommendation module is used to identify plant species and growth needs by taking plant images, and to provide disease and pest diagnosis and pesticide recommendations. The monitoring center module is connected to the maintenance execution module and the community interaction module, the maintenance execution module is connected to the knowledge recommendation module and the spatial adaptation module, and the knowledge recommendation module is connected to the spatial adaptation module.

6. An automated intelligent flowerpot with water monitoring and early warning function according to claim 5, characterized in that, The Q-learning algorithm is used to construct the decision core, discretizing the action space into multiple irrigation amounts and interval durations. A reward function is established by integrating growth indicators and resource efficiency, and a water-saving penalty mechanism is introduced. If an action triggers an interception by the rule engine, a penalty value is applied. Iterative optimization is achieved through continuous interaction, and this process also includes: The system acquires plant images taken by the user and plant species identification results, extracts the morphological features of the plant, and combines them with the decision tree rule base to determine the current growth stage of the plant. The morphological features include the number of leaves, leaf area index, new leaf germination rate, and flower bud ratio. Based on the analysis results of growth stages, a state space for reinforcement learning is constructed with growth stage categorical variables as discrete variables and soil moisture gradient values ​​as continuous variables. The hard constraints of the rule engine are activated to ensure that soil moisture is always maintained within the safe humidity range for plants.

7. An automated intelligent flowerpot with water monitoring and early warning function according to claim 5, characterized in that, Establish a user-generated content community, generate a maintenance index based on maintenance operations, support users in publishing maintenance logs and unlocking achievements, and automatically group users based on their habits, including: Real-time recording of maintenance operations generates four-dimensional maintenance indicators. A weighted algorithm generates a user-specific maintenance index. Based on a pre-built two-layer community structure, a dual-track driving mechanism combining explicit and implicit incentives is set up to enable user communication and interaction within the user-generated content community. The system obtains the types of plant species that users grow, combines them with user behavior characteristics, uses cosine similarity to measure user distance, matches similar plant planting groups, and supports users in publishing and sharing maintenance logs. Among these user behavior characteristics are user operation time preferences, device usage characteristics, and content interaction patterns.

8. An automated intelligent flowerpot with water monitoring and early warning function according to claim 7, characterized in that, The two-tiered community structure includes a basic layer and a relational layer. The basic layer provides an interface for publishing maintenance logs in text, image, and video formats, while the relational layer constructs an interest tag tree based on user geographic location and plant species tags. The four-dimensional maintenance indicators include the frequency of integrated operations, the diversity of operation types, the improvement value of plant health, and the degree of environmental compatibility. Explicit incentives include progress-based, skill-based, and contribution-based incentives; The implicit incentive uses contribution value decay, which is calculated by reducing the contribution value based on the user's inactive time, and different levels of activity are divided according to the contribution value.

9. An automated intelligent flowerpot with water monitoring and early warning function according to claim 5, characterized in that, Using augmented reality technology to build environmental models, recommending suitable plants, simulating and predicting future plant growth, and supporting interactive adjustments, including: Using a mobile terminal to photograph the planting area where the flowerpot is located, environmental perception is achieved. The object boundary coordinates are calculated using the principle of triangulation, and a spatial mapping model is constructed. The environmental parameters of the current planting area are matched with the plant growth requirements, and a list of recommended plant species is generated by prioritizing the requirements. The virtual plant model is then loaded into the mobile terminal. Load the pre-set flowerpot main model and plant growth keyframe model, use a linear interpolation algorithm to generate a continuous growth animation, map it to the spatial mapping model, and observe whether the plant conflicts with the physical space. If there is a conflict, it is not recommended to plant in the current planting area.

Images