Intelligent farmland irrigation device

By using a sliding gate with an arc trajectory and a cloud-based intelligent irrigation management system, the problems of high gate friction and insufficient intelligence in existing farmland irrigation devices have been solved, realizing intelligent irrigation and remote control, reducing motor power requirements and minimizing manual intervention.

CN118947515AInactive Publication Date: 2024-11-15NINGBO INST OF TECH ZHEJIANG UNIV ZHEJIANG

Patent Information

Application Number
CN202410986910.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-23
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Existing farmland irrigation devices suffer from problems such as high friction between the gate and the gate frame, high power motor requirements, and insufficient intelligence, making it impossible to automatically adjust the water level according to the rice growing season.

Method used

The system employs a sliding gate with an arc trajectory and a cloud-based intelligent irrigation management system. It consists of components such as a small weather station, pumping station, inlet gate, outlet gate, rice growth monitor, cloud server, and mobile APP. Combined with a deep learning neural network model and MQTT server, it realizes automatic gate control and intelligent irrigation.

Benefits of technology

It enables remote control of the gate and intelligent irrigation, reduces the power requirements of the drive motor, avoids a decrease in the gate's sealing performance, reduces reliance on manual labor, and improves the automation level of irrigation management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118947515A_ABST
    Figure CN118947515A_ABST
Patent Text Reader

Abstract

An intelligent farmland irrigation device is composed of a small meteorological station, a pump station, a water inlet gate, a water outlet gate, a rice growth condition monitor, a cloud server and a mobile phone APP. The water inlet gate adopts a translation gate with an arc-shaped track; the rice growth condition monitor collects a growth image of rice and distinguishes the growth period of the rice; in the cloud server, data such as the highest temperature, the lowest temperature, the air humidity, the rainfall, the wind direction, the rice growth period and the soil humidity of each day serve as input quantities, the water requirement of a farmland serves as an output quantity, and a deep learning neural network model of rice irrigation management is obtained by training historical data. On-site environment data and rice growth period estimation are input into the model, the water demand of the farmland is calculated, and automatic irrigation control is carried out by comparing the water demand with a liquid level sensing acquisition value of the paddy field. The gate can be remotely controlled, irrigation intellectualization can be achieved, and dependence of irrigation management on manpower is reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the field of farmland water conservancy equipment, and in particular relates to farmland irrigation equipment. Background Art

[0002] Through the construction of farmland water conservancy facilities, my country has built a criss-cross hardened canal irrigation system in the southern rice-producing areas, ensuring the stable and increased production of rice. These hardened canals will reserve a gate at the edge of each farmland. When people pull out the wooden board in the gate, they will start to irrigate the farmland. When the wooden board is inserted into the gate, the irrigation will stop, so that the water continues to flow along the canal to prepare for irrigating other farmlands. In the hot summer, rice needs a lot of water and needs to be irrigated frequently; the water demand and water level requirements are different in different growth stages of rice. Sometimes drainage is even required to promote rice growth, so it is necessary to inspect the farmland every day. It is estimated that the maximum irrigation management capacity of an artificial person is only 500 acres / day. In the situation of increasing labor shortage, automatic and intelligent irrigation equipment is urgently needed. Some new electric gates have been disclosed. For example, the utility model patent "An electric gate" (patent number CN201720277067) proposes a structure in which a linear drive controls the rise or fall of the gate plate. The advantage is that the purpose of electrically controlling the gate is achieved. However, when the gate is lifted or lowered, due to the water pressure, there is a large friction between the gate plate and the gate frame, so a high-power motor is required to provide power. Another example is the invention patent "A double-guide rod electric gate structure" (CN201720006554). The rope pulley guide rod mechanism of the double-guide rod electric gate structure is a rolling friction mechanism with high transmission efficiency, but there is still sliding friction between the gate and the column. It has the same disadvantage as the previous patent and still requires a high-power motor. There is also an invention patent "A channel electric gate" (CN202220254050). This patent uses a driver to electronically control the gate plate to lift and close. It is easy to operate, saves time and effort, but the gate plate and the column in this patent also have sliding friction, which is consistent with the shortcomings of the two patents mentioned above. Both require a high-power motor. The invention patent application (202310894998.3) uses wireless communication and liquid level sensors to achieve automatic control of the water level, but it is also necessary to manually set the irrigation water level according to different rice growth periods. In terms of intelligence, it needs to be further combined with agronomic requirements. Moreover, the wireless communication method also affects the convenience and cost of use. Summary of the invention

[0003] In order to overcome the defects of existing farmland irrigation devices, the present invention provides a sliding gate with an arc track and a cloud-based intelligent irrigation management system. The sliding gate with an arc track can not only completely avoid the friction between the gate and the gate frame during the opening and closing process, reduce the power of the driving motor, and avoid the reduction of the gate sealing due to wear; by collecting rice images and growth environment data, the deep learning neural network model is used in the cloud to calculate the water demand of rice, which is automatically executed by the gate, minimizing the dependence of field irrigation on manual labor.

[0004] The technical solution adopted by the present invention to solve its technical problems is: the farmland intelligent irrigation device consists of 7 independent components, including a small weather station, a pump station, a water inlet gate, a drainage gate, a rice growth period monitor, a cloud server and a mobile phone APP, wherein the small weather station consists of a shutter box, a weather station irrigation control panel, an air temperature and humidity sensor, a rainfall sensor, a light sensor and a wind speed sensor; the pump station consists of a submersible pump, a pump station irrigation control panel, a river liquid level sensor and a canal liquid level sensor; the mechanical structure of the water inlet gate consists of a door frame, a gate, a sealing ring, an upper rocker, a lower rocker, an electric push rod, a bracket, a rotating shaft, a coupling and an angular displacement sensor. The middle part of the door frame of the water inlet gate is a hollow rectangular water hole; a sealing ring is pasted at each contact point between the gate and the door frame; the upper and lower hinge supports, upper rocker, lower rocker and gate of the door frame constitute the parallelogram motion mechanism of the gate, the center of the gate is a protruding hinge support, the extended end of the electric push rod is hinged to the hinge support, and the body of the electric push rod is hinged to the support; when the electric push rod pushes or pulls, the gate is subject to the geometric constraints of the parallelogram and moves in an arc trajectory. When a pair of sealing rings fit and press, the water inlet gate is closed. When the gate and the door frame are separated from each other, water flows into the water hole from the bottom, left and right sides of the gate to start irrigating the farmland. The rotating shaft and the upper rocker are welded as one. The movement of the gate is transmitted to the rotating shaft through the upper rocker. The end of the rotating shaft is connected to the angular displacement sensor through a coupling. This angular displacement sensor measures the opening of the gate. The stroke of the electric push rod is 50mm. The center hole spacing of the upper rocker and the lower rocker is 55mm, which ensures that the maximum rotation angle of the upper rocker and the lower rocker is close to but not more than 90°. This parallelogram motion mechanism does not have a dead angle, will not get stuck, and will not pull the door frame when the stroke of the electric push rod reaches the maximum or minimum. The drainage gate consists of an electric push rod, a ball valve, an annular sealing ring, a drainage basin seat, a drain outlet, and a bracket. The diameter of the ball valve is 150mm. Its upper handle is connected to the electric push rod, and the ball valve moves in the vertical direction. The body of the electric push rod is fixed on the bracket fixed to the surface of the drainage basin seat. When the electric push rod is extended, the ball valve is pressed down until the ball valve is pressed tightly against the annular sealing ring attached to the drainage pipe seat to close the drainage. Otherwise, the ball valve is opened to drain.

[0005] The water inlet gate measurement and control device includes an irrigation control board, an antenna, a solar panel, a charger, a battery, a gate opening and closing button, a farmland liquid level sensor, a canal liquid level sensor, a soil moisture sensor, an angular displacement sensor and a relay module. The pin 1 of each gate opening and closing button is grounded, and the pin 2 is connected to the IN pin of the relay module; the irrigation control board is composed of an ESP32 module, a signal excitation and acquisition circuit, a filter circuit, an output circuit, an RS485 circuit and a wireless data transmission module; the ESP32 module has a built-in WiFi antenna. When the software is started, if there is a WiFi signal, it connects to the MQTT server, otherwise it communicates data with the remote weather station terminal through the wireless data transmission module, and the weather station terminal is connected to the MQTT server; the 2 outputs of the ESP32 module are connected to the IN pin of the relay module through the output circuit, so that the water inlet gate can be opened and closed by the program, or the gate opening and closing button can be directly operated to open and close the water inlet gate. The farmland liquid level sensor consists of a liquid level electrode inserted in the water and a signal excitation and acquisition circuit on the irrigation control board. The liquid level electrode is a densely interdigitated capacitor made in the form of a PCB with an insulating coating on the surface. The electrode is fixed to the side of a wooden frame with a conical head. The wooden frame is inserted into the soil so that the bottom of the electrode is flush with the soil surface. The signal excitation and acquisition circuit generates a square wave excitation signal through an external resistor and capacitor charging and discharging circuit by a 555 timer chip. This signal passes through the first 1uf capacitor and outputs a pure AC square wave signal; the AC square wave signal is input into a resistor-capacitor series circuit, and the positive terminal of the capacitor generates a triangle wave signal. This triangle wave signal flows into the first-stage operational amplifier to form a sine wave signal, and then passes through the second 1uf capacitor to output a pure AC sine wave signal. This signal is input into the resistor-capacitor series circuit composed of a capacitor and a resistor formed by the liquid level electrode, and an AC signal reflecting the liquid level is output from the positive terminal of the resistor. It passes through the second-stage operational amplifier to form a voltage follower, and then flows into the third 1uf capacitor. The output signal passes through an AC-to-DC signal circuit composed of a diode and a capacitor. This liquid level signal is input into the filter circuit, collected by the esp32 module, and constitutes liquid level data, which is published to the MQTT server through wireless communication. The canal liquid level sensor is inserted into the canal, and also obtains the excitation signal from the irrigation control board, which is collected by the esp32 module to constitute liquid level data, which is published to the MQTT server through wireless communication. The soil moisture sensor is inserted into the soil next to the farmland liquid level sensor, and sends the current soil moisture to the esp32 module through the RS485 circuit. After the irrigation control board obtains the soil moisture data, it is published to the MQTT server through wireless communication. As long as the mobile phone APP subscribes to the information of the angular displacement sensor, liquid level sensor and soil moisture sensor, it can regularly obtain the real-time status of the water inlet, the farmland water level, the canal water level and the moisture of the farmland soil, and can also remotely open and close the water inlet.

[0006] After the irrigation control board of the water inlet gate is powered on, it generates an electric push rod object, an analog signal object, a USART object, a timer object, an MQTT object and a WiFi object. The MQTT object subscribes to the "remote control" message. When the subscribed "remote control" message arrives, the irrigation control board identifies the instruction and the electric push rod object opens and closes the water inlet gate; when the timing moment of the timer object is triggered, the analog signal object is immediately used to collect the signals of the angular displacement sensor, the farmland liquid level sensor, and the canal liquid level sensor, and the soil moisture sensor is requested for data through the USART object. At this time, if the liquid level of the farmland is lower than the set minimum water level threshold, the electric push rod object is immediately used to open the water inlet gate, and the "remote control" message of opening the pump station is published to the wireless data transmission module or the MQTT object through the USART object, so that the canal can continuously obtain irrigation water; when the collected value of the farmland liquid level sensor (18) is higher than the maximum water level threshold, the electric push rod object is immediately used to close the water inlet gate. When the soil moisture message of the USART is received, the liquid level and soil moisture information are published to the MQTT server through the USART object to the wireless data transmission module or the MQTT object.

[0007] After the irrigation control board of the drainage gate is powered on, it generates an electric push rod object, an analog signal object, a USART object, a timer object, an MQTT object and a WiFi object. The MQTT object subscribes to the "remote control" message. When the subscribed "remote control" message arrives, the irrigation control board identifies the command and the electric push rod object opens and closes the drainage gate; when the timing moment of the timer object is triggered, the analog signal object is used to collect the signals of the farmland liquid level sensor and the canal liquid level sensor, and the liquid level information is published to the MQTT server through the USART object to the wireless data transmission module or the MQTT object. In the water storage state, when the irrigation control board detects that the liquid level drop rate of the farmland exceeds the natural evaporation drop rate, an alarm message of drainage gate leakage is immediately issued.

[0008] After the irrigation control board of the pump station is powered on, it generates a submersible pump object, an analog signal object, a timer object, a USART object, an MQTT object, and a WiFi object. The MQTT object subscribes to the "remote control" message. When the subscribed "remote control" message arrives, the irrigation control board identifies the instruction and the submersible pump object starts and closes the submersible pump; when the timing moment of the timer object is triggered, the analog signal object is used to collect the signal of the river level sensor and the canal level sensor, and the liquid level information is published to the MQTT server through the USART object to the wireless data transmission module or the MQTT object.

[0009] After the irrigation control board of the weather station is powered on, it generates USART objects, timer objects, MQTT objects and WiFi objects. The MQTT object subscribes to the "remote control" message. When the subscribed "remote control" message arrives, the irrigation control board identifies the target terminal address of the forwarded message and sends the "remote control" message to the wireless data transmission module through the USART object; when the timer object is triggered, it immediately requests data from the air temperature and humidity, rainfall, illuminance and wind speed sensors through the USART object; when the USART object receives the message, the MQTT object publishes the meteorological information to the MQTT server.

[0010] The rice growth period monitor uses K210 as the core board, continuously monitors 50 points, and collects rice pictures once every day from transplanting to harvesting, forming the initial 200 training data sets and test data sets, which are input to the Maxihub website for online annotation, marking the rice's greening period, tillering period, panicle development period, heading and fruiting period, and performing deep learning online training to generate a rice growth period model, which is downloaded to the K210 chip. In the actual rice growth environment, the rice growth period monitor regularly collects rice pictures, inputs them into the rice growth period model, outputs the judgment of the rice growth stage, and publishes the liquid level information to the MQTT server through the USART object to the wireless data transmission module or the MQTT object. After the intelligent irrigation system is running, it continues to accumulate and annotate the monitored pictures every day, conducts deep learning online training of new and old data sets in each production season, and updates the rice growth period model to make this model more and more accurate.

[0011] The cloud server builds a rice intelligent irrigation management model. For different rice varieties, during the rice planting period, the daily maximum temperature, minimum temperature, air humidity, rainfall, wind direction, rice growth period and soil humidity are used as input, and the farmland water requirements - drained (water level is 0, soil moisture is between 80%-95%), shallow wet dry (water level is 0-3cm), low water level (water level is 3-6cm), high water level (water level is 6-10cm) are used as outputs. A deep learning model with 7 inputs, 2 hidden layers, 32 nodes in each layer and 4 outputs is constructed. The historical agricultural production data is divided into training data and test data, and a deep learning neural network model for rice irrigation management is obtained through training. In the actual environment of rice growth, the on-site production data is input into this model every 6 hours to calculate the water demand of the farmland, generate the low and high water levels of the farmland water level, and publish it by the MQTT server. After the intelligent irrigation system is in operation, the production data and water level measurements of the farmland required by this model are still accumulated every day. Deep learning online training is carried out on new and old data sets every production season to update the intelligent irrigation management model and make it more and more accurate.

[0012] The beneficial effects of the present invention are that not only the gate can be remotely controlled, but also intelligent irrigation can be realized, completely liberating labor. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 It is an embodiment of the mechanical structure of the water intake gate.

[0014] Figure 2 It is an embodiment of the mechanical structure of the drainage gate.

[0015] Figure 3 The invention is an embodiment of a water inlet gate measuring and controlling device.

[0016] Figure 4 It is a software flow chart embodiment of an irrigation control panel for a water intake gate.

[0017] Figure 5 It is a software flow chart embodiment of an irrigation control panel for a drainage gate.

[0018] Figure 6 It is a software flow chart embodiment of an irrigation control panel of a pump station.

[0019] Figure 7 It is a software flow chart embodiment of an irrigation observation board of a weather station.

[0020] Figure 8 The invention is an embodiment of the operation flow of the rice growth period monitor to form a rice growth period model.

[0021] Fig. 9 It is an embodiment of the operation flow of forming a rice intelligent irrigation management model during the rice growing period.

[0022] As shown in the figure, 1. door frame, 2. upper rocker, 3. bracket, 4. electric push rod, 5. sealing ring, 6. gate, 7. sealing ring, 8. lower rocker, 9. angular displacement sensor, 10. coupling, 11. rotating shaft, 12. bracket, 13. electric push rod, 14 drainage basin seat, 15. ball valve, 16. annular sealing ring, 17. drain outlet, 18. farmland liquid level sensor, 19. canal liquid level sensor, 20. irrigation control board, 21. solar power panel, 22. charger, 23. battery, 24. antenna, 25. gate opening and closing button, 26. relay module, 27. soil moisture sensor. Specific implementation plan The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments: Figure 1The invention is an embodiment of the mechanical structure of a water inlet gate, which comprises a door frame (1), a gate (6), a sealing ring (5, 7), an upper rocker (2), a lower rocker (8), an electric push rod (4), a bracket (3), a rotating shaft (11), a coupling (10), and an angular displacement sensor (9). The middle part of the door frame (1) of the water inlet gate is a hollow rectangular water hole; a sealing ring (5, 7) is pasted at each contact point between the gate (6) and the door frame (1); the upper and lower hinge supports of the door frame (1), the upper rocker (2), the lower rocker (8) and the gate (6) constitute a parallelogram motion mechanism of the gate, the central part of the gate (6) is a protruding hinge support, the extended end of the electric push rod (4) is hinged to the hinge support, and the body of the electric push rod (4) is hingedly connected to the support (3); when the electric push rod (4) pushes or pulls, the gate (6) is constrained by the parallelogram geometry to move in an arc trajectory, when a pair of sealing rings (5, 7) fit and press, the water inlet gate is closed, when the gate (6) and the door frame (1) are separated from each other, water flows from the lower side, left side and right side of the gate (6) into the water hole, and begins to irrigate the farmland. The rotating shaft (11) and the upper rocker (2) are welded as one body, the movement of the gate (6) is transmitted to the rotating shaft (11) through the upper rocker (2), and the end of the rotating shaft (11) is connected to the angular displacement sensor (9) through a coupling (10), and the angular displacement sensor (9) measures the opening of the gate (6); the stroke of the electric push rod (4) is 50 mm, and the center hole spacing between the upper rocker (2) and the lower rocker (8) is 55 mm, ensuring that the maximum rotation angle of the upper rocker (2) and the lower rocker (8) is close to but not more than 90°. This parallelogram motion mechanism does not have a dead angle, will not get stuck, and will not pull the door frame (1) when the stroke of the electric push rod (4) reaches the maximum or minimum.

[0024] Figure 2 The invention is an embodiment of the mechanical structure of a drainage gate. The drainage gate is composed of an electric push rod (13), a ball valve (15), an annular sealing ring (16), a drainage basin seat (14), a water outlet (17), and a bracket (12). The diameter of the ball valve (15) is 150 mm. Its upper handle is connected to the electric push rod (13). The ball valve (15) moves in the vertical direction. The body of the electric push rod (13) is fixed on the bracket (12) fixed to the surface of the drainage basin seat (14). When the electric push rod (13) is extended, the ball valve (15) is pressed downward until the ball valve (15) is pressed against the annular sealing ring (16) attached to the drainage pipe seat (14) to close the drainage. Otherwise, the ball valve (15) is opened to drain the water.

[0025] Figure 3The invention is an embodiment of a water inlet gate measurement and control device. The water inlet gate measurement and control device comprises an irrigation control panel (20), an antenna 24 (), a solar power generation panel (21), a charger (22), a storage battery (23), a gate opening and closing button (25), a farmland liquid level sensor (18), a canal liquid level sensor (19), a soil moisture sensor (27), an angular displacement sensor (9) and a relay module (26). The pin 1 of each gate opening and closing button (25) is grounded, and the pin 2 is connected to the IN pin of the relay module (26); the irrigation control board (20) is composed of an ESP32 module, a signal excitation and acquisition circuit, a filtering circuit, an output circuit, an RS485 circuit and a wireless data transmission module; the ESP32 module has a built-in WiFi antenna. When the software is started, if there is a WiFi signal, it connects to the MQTT server, otherwise it communicates data with the remote weather station terminal through the wireless data transmission module, and the weather station terminal is connected to the MQTT server; the two outputs of the ESP32 module are respectively connected to the IN pin of the relay module (26) through the output circuit, so that the water inlet gate can be opened and closed by the program, and the water inlet gate can also be opened and closed by directly operating the gate opening and closing button (25). The farmland liquid level sensor (18) is composed of a liquid level electrode inserted into water and a signal excitation and acquisition circuit on an irrigation control board (20). The liquid level electrode is a densely interdigitated capacitor made in the form of a PCB, and the surface is covered with an insulating coating. The electrode is fixed to the side of a wooden frame with a conical head, and the wooden frame is inserted into the soil so that the bottom of the electrode is flush with the soil surface. The signal excitation and acquisition circuit generates a square wave excitation signal through an external resistor and capacitor charging and discharging circuit by a 555 timer chip. This signal passes through the first 1uf capacitor to output a pure AC square wave signal; the AC square wave signal is input into a resistor-capacitor series circuit, and the positive terminal of the capacitor generates a triangle wave signal. This triangle wave signal flows into the first-stage operational amplifier to form a sine wave signal, and then passes through the second 1uf capacitor to output a pure AC sine wave signal. This signal is input into the resistor-capacitor series circuit composed of a capacitor and a resistor formed by the liquid level electrode, and an AC signal reflecting the liquid level is output from the positive terminal of the resistor. It passes through the second-stage operational amplifier to form a voltage follower, and then flows into the third 1uf capacitor. The output signal passes through an AC-to-DC signal circuit composed of a diode and a capacitor. The liquid level signal is input into the filter circuit, collected by the esp32 module, constituted as liquid level data, and published to the MQTT server via wireless communication; the canal liquid level sensor (19) is inserted into the canal, and also obtains an excitation signal from the irrigation control board (20), which is collected by the esp32 module, constituted as liquid level data, and published to the MQTT server via wireless communication; the soil moisture sensor (27) is inserted into the soil next to the farmland liquid level sensor, and sends the current soil moisture to the esp32 module via the RS485 circuit. After the irrigation control board (20) obtains the soil moisture data, it is published to the MQTT server via wireless communication.As long as the mobile phone APP subscribes to the information of the angular displacement sensor, liquid level sensor and soil moisture sensor, it can regularly obtain the real-time status of the water inlet gate, farmland water level, canal water level and farmland soil humidity, and can also remotely open and close the water inlet gate.

[0026] Figure 4 The invention is a software flow chart embodiment of an irrigation control board (20) for a water inlet gate. After the irrigation control board (20) for the water inlet gate is powered on, it generates an electric push rod object, an analog signal object, a USART object, a timer object, an MQTT object and a WiFi object, and the MQTT object subscribes to a "remote control" message. When the subscribed "remote control" message arrives, the irrigation control board (20) identifies the instruction and opens and closes the water inlet gate by the electric push rod object; when the timing moment of the timer object is triggered, the analog signal object is used to collect signals from the angular displacement sensor (9), the farmland liquid level sensor (18) and the canal liquid level sensor (19), and data is requested from the soil moisture sensor (27) through the USART object. At this time, if the liquid level in the farmland is lower than the set minimum water level threshold, the electric push rod object will immediately open the water inlet gate, and publish a "remote control" message to open the pump station to the wireless data transmission module or MQTT object through the USART object, so that the canal can continue to obtain irrigation water; when the collected value of the farmland liquid level sensor (18) is higher than the maximum water level threshold, the electric push rod object will immediately close the water inlet gate. When the soil moisture message from the USART is received, the liquid level and soil moisture information will be published to the MQTT server through the USART object to the wireless data transmission module or MQTT object.

[0027] Figure 5 The invention is a software flow chart embodiment of an irrigation control board (20) for a drainage gate. After the irrigation control board (20) for the drainage gate is powered on, it generates an electric push rod object, an analog signal object, a USART object, a timer object, an MQTT object and a WiFi object, and the MQTT object subscribes to a "remote control" message. When the subscribed "remote control" message arrives, the irrigation control board (20) identifies the instruction and opens and closes the drainage gate by the electric push rod object; when the timing moment of the timer object is triggered, the analog signal object is used to collect the signal of the farmland liquid level sensor (18) and the canal liquid level sensor (19), and the liquid level information is published to the MQTT server through the USART object to the wireless data transmission module or the MQTT object. In the water storage state, when the irrigation control board (20) detects that the liquid level drop rate of the farmland exceeds the natural evaporation drop rate, an alarm message of drainage gate leakage is immediately published.

[0028] Figure 6The invention is a software flow chart embodiment of an irrigation control board (20) of a pump station. After the irrigation control board (20) of the pump station is powered on, it generates a submersible pump object, an analog signal object, a timer object, a USART object, an MQTT object and a WiFi object, and the MQTT object subscribes to a "remote control" message. When the subscribed "remote control" message arrives, the irrigation control board (20) identifies the instruction, and the submersible pump object starts and closes the submersible pump; when the timing moment of the timer object is triggered, the analog signal object is used to collect the signal of the river liquid level sensor and the canal liquid level sensor (19), and the liquid level information is published to the MQTT server through the USART object to the wireless data transmission module or the MQTT object.

[0029] Figure 7 The invention is a software flow chart embodiment of an irrigation control board (20) of a weather station. After the irrigation control board (20) of the weather station is powered on, it generates a USART object, a timer object, an MQTT object and a WiFi object, and the MQTT object subscribes to a "remote control" message. When the subscribed "remote control" message arrives, the irrigation control board (20) identifies the target terminal address of the forwarded message and sends the "remote control" message to the wireless data transmission module through the USART object; when the timing moment of the timer object is triggered, the USART object is immediately used to request data from the air temperature and humidity, rainfall, illuminance and wind speed sensor; when the USART object receives the message, the MQTT object publishes the meteorological information to the MQTT server.

[0030] Figure 8 This is an embodiment of the operation flow of the rice growth period monitor to form a rice growth period model. The rice growth period monitor uses K210 as the core board, continuously monitors 50 points, and collects rice pictures once every day from transplanting to harvesting, forming the initial 200 sets of training data sets and test data sets, which are input to the Maxihub website for online annotation, marking the rice's greening period, tillering period, panicle development period, and heading and fruiting period, and performing deep learning online training to generate a rice growth period model, which is downloaded to the K210 chip. In the actual environment where rice grows, the rice growth period monitor regularly collects rice pictures, inputs them into the rice growth period model, and outputs the judgment of the rice growth stage. After the intelligent irrigation system is running, it continues to accumulate monitored pictures and annotate them every day, and conducts deep learning online training of new and old data sets in each production season to update the rice growth period model, making this model more and more accurate.

[0031] Fig. 9This is an embodiment of the operation flow of forming a rice intelligent irrigation management model during the rice growing period. On the cloud server, for different rice varieties, during the rice planting period, the highest temperature, lowest temperature, air humidity, rainfall, wind direction, rice growing period and soil humidity of each day are used as inputs, and the farmland water demand requirements - drained (water level is 0, soil moisture is between 80%-95%), shallow wet dry (water level is 0-3cm), low water level (water level is 3-6cm), high water level (water level is 6-10cm) are used as outputs. A deep learning model with 7 inputs, 2 hidden layers, 32 nodes per layer, and 4 outputs is constructed. The historical agricultural production data is divided into training data and test data, and a deep learning neural network model for rice irrigation management is obtained through training; in the actual environment of rice growth, every 6 hours, the field production data and rice growing period are input into this model, the water demand of the farmland is calculated, and the low and high water levels of the farmland water level are generated, which are published by the MQTT server. Similarly, after the smart irrigation system is in operation, the production data and water level measurements of the farmland required by this model are still accumulated every day. Deep learning online training is carried out on new and old data sets every production season to update the smart irrigation management model and make it more and more accurate.

Claims

1. A smart farm irrigation device, consisting of seven independent components: a small weather station, a pump station, a water inlet gate, a water outlet gate, a rice growth period monitor, a cloud server and a mobile phone APP, characterized in that: The cloud server builds a smart rice irrigation management model. For different rice varieties, during the rice planting period, the production data such as the highest temperature, lowest temperature, air humidity, rainfall, wind direction, rice growth period and soil humidity of each day are used as input, and the farmland water demand is used as output. A deep learning model with 7 inputs, 2 hidden layers, 32 nodes in each layer and 4 outputs is built. The historical agricultural production data is divided into training data and test data, and the deep learning neural network model of rice irrigation management is obtained through training. In the actual environment of rice growth, the on-site production data is input into this model every 6 hours to calculate the water demand of the farmland, generate the low and high water levels of the farmland water level, and publish them by the MQTT server. After the smart irrigation system is running, the production data and farmland water level measurements required by this model are still accumulated every day. Deep learning online training of new and old data sets is carried out in each production season to update the smart irrigation management model.

2. The intelligent farmland irrigation device according to claim 1 is characterized in that: The middle part of the door frame of the water inlet gate is a hollow rectangular water hole; a sealing ring is pasted at each contact point between the gate and the door frame; the upper and lower hinge supports, upper rocker, lower rocker and gate of the door frame constitute the parallelogram motion mechanism of the gate, the center of the gate is a protruding hinge support, the extended end of the electric push rod is hinged to the hinge support, and the body of the electric push rod is hinged to the support; when the electric push rod pushes or pulls, the gate is subject to the geometric constraints of the parallelogram and moves in an arc trajectory. When a pair of sealing rings fit and press, the water inlet gate is closed. When the gate and the door frame are separated from each other, water flows into the water hole from the bottom, left and right sides of the gate to start irrigating the farmland. The rotating shaft and the upper rocker are welded as one, and the movement of the gate is transmitted to the rotating shaft through the upper rocker. The end of the rotating shaft is connected to the angular displacement sensor through a coupling. This angular displacement sensor measures the opening of the gate; the stroke of the electric push rod is 50mm, and the distance between the center holes of the upper rocker and the lower rocker is 55mm, which ensures that the maximum rotation angle of the upper rocker and the lower rocker is close to but not more than 90°. This parallelogram motion mechanism has no dead angle, will not get stuck, and will not pull the door frame when the stroke of the electric push rod reaches the maximum or minimum.

3. The intelligent farmland irrigation device according to claim 1 is characterized in that: The diameter of the ball valve of the drainage gate is 150mm. Its upper handle is connected to the electric push rod, and the ball valve moves in the vertical direction. The body of the electric push rod is fixed on a bracket fixed to the surface of the drainage basin seat. When the electric push rod is extended, the ball valve is pressed downward until the ball valve is pressed tightly against the annular sealing ring on the drainage pipe seat to close the drainage. Otherwise, the ball valve is opened to drain the water.

4. The intelligent farmland irrigation device according to claim 1 is characterized in that: The pin 1 of each gate opening and closing button of the water inlet measurement and control device is grounded, and the pin 2 is connected to the IN pin of the relay module; the ESP32 module of the irrigation control board has its own WiFi antenna. When the software is started, if there is a WiFi signal, it will connect to the MQTT server, otherwise it will communicate data with the remote weather station terminal through the wireless data transmission module, and the weather station terminal will be connected to the MQTT server; the 2 outputs of the ESP32 module are respectively connected to the IN pin of the relay module through the output circuit, so that the water inlet can be opened and closed by the program, or the gate opening and closing button can be directly operated to open and close the water inlet; the farmland liquid level sensor consists of a liquid level electrode inserted in the water and a signal excitation and acquisition circuit on the irrigation control board. The liquid level electrode is a dense forked-finger capacitor made in PCB form with an insulating coating on the surface. The electrode is fixed to the side of a wooden frame with a conical head. The wooden frame is inserted into the soil so that the bottom of the electrode is flush with the soil surface; signal excitation and acquisition circuit The formed pure AC sine wave signal is input into the RC series circuit composed of the capacitor and resistor formed by the liquid level electrode, and the AC signal reflecting the liquid level is output from the positive end of the resistor. It passes through the operational amplifier to form a voltage follower, and then flows into a 1uf capacitor. The output signal passes through the AC-DC signal circuit composed of a diode and a capacitor; this liquid level signal is input into the filter circuit, collected by the esp32 module, constitutes liquid level data, and is published to the MQTT server through wireless communication; the canal liquid level sensor is inserted into the canal, and also obtains the excitation signal from the irrigation control board, which is collected by the esp32 module to constitute liquid level data, and is published to the MQTT server through wireless communication; the soil moisture sensor is inserted into the soil next to the farmland liquid level sensor, and sends the current soil moisture to the esp32 module through the RS485 circuit. After the irrigation control board obtains the soil moisture data, it is published to the MQTT server through wireless communication.

5. The intelligent farmland irrigation device according to claim 1 is characterized in that: After the irrigation control board of the water inlet gate is powered on, it generates an electric push rod object, an analog signal object, a USART object, a timer object, an MQTT object and a WiFi object. The MQTT object subscribes to a "remote control" message. When the subscribed "remote control" message arrives, the irrigation control board identifies the instruction and the electric push rod object opens and closes the water inlet gate. When the timing moment of the timer object is triggered, the analog signal object is used to collect signals from the angular displacement sensor, the farmland liquid level sensor and the canal liquid level sensor, and the soil moisture sensor is requested for data through the USART object. At this time, if the liquid level of the farmland is lower than the set minimum water level threshold, the electric push rod object is used to open the water inlet gate immediately, and a "remote control" message for opening the pump station is published to the wireless data transmission module or the MQTT object through the USART object, so that the canal can continuously obtain irrigation water. When the collected value of the farmland liquid level sensor (18) is higher than the maximum water level threshold, the electric push rod object is used to close the water inlet gate immediately. When the soil moisture message of the USART is received, the liquid level and soil moisture information are published to the MQTT server through the USART object to the wireless data transmission module or the MQTT object.

6. The intelligent farmland irrigation device according to claim 1 is characterized in that: When the subscribed "remote control" message arrives at the irrigation control board of the weather station, the irrigation control board identifies the target terminal address of the forwarded message and sends the "remote control" message to the wireless data transmission module through the USART object; when the timing moment of the timer object is triggered, the air temperature and humidity, rainfall, light intensity and wind speed sensors are immediately requested for data through the USART object; when the USART object receives the message, the MQTT object publishes the meteorological information to the MQTT server.

7. The intelligent farmland irrigation device according to claim 1 is characterized in that: The cloud server builds a rice intelligent irrigation management model. For different rice varieties, during the rice planting period, the production data of the highest and lowest temperatures, air humidity, rainfall, wind direction, rice growth period and soil humidity of each day are used as input, and the farmland water requirements - drained (water level is 0, soil moisture is between 80%-95%), shallow wet dry (water level is 0-3cm), low water level (water level is 3-6cm), high water level (water level is 6-10cm) are used as output. A 7-input, 2-hidden layer, 32 nodes per layer, and 4 output nodes are constructed. The deep learning model divides the historical agricultural production data into training data and test data, and obtains the deep learning neural network model of rice irrigation management through training; in the actual environment of rice growth, the on-site production data is input into this model every 6 hours to calculate the water demand of the farmland, generate the low and high water levels of the farmland, and publish them by the MQTT server; after the intelligent irrigation system is running, the production data and farmland water level measurements required by this model are still continuously accumulated every day, and deep learning online training of new and old data sets is carried out in each production season to update the intelligent irrigation management model.

Citation Information

Patent Citations

  • Intelligent irrigation and drainage regulation and control system and method for rice field

    CN116746466A

  • Two guide arm electric valve structures

    CN206368346U

  • Electric valve

    CN206859189U

  • Electric gate for channel

    CN217267339U

Cited By

  • Agricultural automatic irrigation system and method based on big data

    CN119678832A