Self-organizing network migratory pest control system based on insect radar and edge computing

Abstract

The present invention relates to the technical field of migratory pest monitoring and prevention and control, and in particular to a self-organizing network migratory pest prevention and control system based on insect radar and edge computing. The technical scheme thereof comprises: linking the flight behavior data of migratory pests monitored by insect radar with a multi-spectral light attractant through radar-controlled light fusion, guiding the dynamic adjustment of the prevention and control strategy, and controlling a plurality of slave prevention and control lights to work collaboratively through a master control light through a self-organizing network protocol, thereby reducing costs and greatly increasing the light attractant coverage area. The edge-cloud collaborative computing architecture is adopted, and the edge computing technology completes pest identification locally. A lightweight insect target detection model is deployed at the light end to realize second-level migratory pest identification and counting. The cloud regularly pushes the latest model weights to improve the identification accuracy, thereby realizing accurate monitoring, real-time identification and three-dimensional interception of migratory pests, solving the problems of delayed response, accidental injury to natural enemies, high energy consumption and limited coverage of traditional light trapping equipment, and significantly improving the prevention and control efficiency of migratory pests.

Description

Technical Field

[0001] The present invention relates to the technical field of migratory pest monitoring and prevention and control, and in particular to a self-organizing network migratory pest prevention and control system based on insect radar and edge computing. Background Art

[0002] Migratory pests are increasingly harmful to agriculture, so it is necessary to monitor and control them. In the existing technology, traditional high-altitude detection lights rely on manual inspections, and cannot obtain pest flight behavior parameters in real time, and the efficiency of light trapping is greatly affected by environmental factors; the existing radar monitoring system is separated from the blocking and control equipment, and the data cannot be coordinated, resulting in high response delays and delayed prevention and control measures, which is difficult to meet the sudden prevention and control needs of migratory pests; the coverage of a single lighting device is limited, and it lacks multi-node collaborative networking capabilities, and cannot form a regional interception network; the existing cloud-based pest identification system is affected by network delays and computing power blockages, and cannot meet the real-time response needs in the field; and turning on the detection lights for a long time is also prone to causing problems such as accidental injury to natural enemies and power consumption.

[0003] In view of this, we propose a self-organizing migratory pest control system based on insect radar and edge computing to solve the existing problems. Summary of the invention

[0004] The purpose of the present invention is to provide a self-organizing network migratory pest control system based on insect radar and edge computing to solve the problems raised in the above background technology.

[0005] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a self-organizing network migratory pest control system based on insect radar and edge computing, including a high-resolution insect radar, a master control lamp, a slave control lamp, and a cloud management platform, wherein the high-resolution insect radar monitors the population of migratory pests in the air and flight parameters in real time, and transmits them to the cloud management platform through the network; the master control lamp includes an edge computing module, a multi-spectral trapping module, an infrared insecticide module, and a communication module; the slave control lamp is connected to the master control lamp through a LoRa wireless link, receives light source control instructions, pitch angle parameters and infrared heating parameters, and feeds back working status data; the cloud management platform receives multi-lamp coordination data through a 4G network, automatically determines the insect peak level according to the real-time monitoring of the insect quantity by the insect radar, calculates the best insect attraction angle according to the insect swarm stratification, flight direction and speed, and remotely and automatically controls the main lamp switch, power and pitch angle through the 4G communication module to achieve the maximum insect attraction;

[0006] Among them, the edge computing module is equipped with a lightweight insect target detection model, with the input being the insect body image collected by the visible light camera and the output being the pest type and density level; the multi-spectral attraction module contains 6 groups of independently modulated LED arrays with a wavelength covering 300-650nm, which adjust the spectrum to adapt to the preferences of different pests, and dynamically combine the light-emitting bands and power ups and downs according to the recognition results; the infrared insecticide module adopts an infrared heating chamber to adjust the heating temperature and heating time according to the type of pest, and adapt to the thresholds of pests of different sizes; the communication module integrates the LoRa SX1276 chip and the 4G Cat.1 module to receive instructions from the insect radar, and realize low-power networking with the slave barrier-controlled lamps and cloud data backhaul.

[0007] Furthermore, the LoRa self-organizing network protocol adopts a star topology structure, with the master resistance-controlled light as the network coordinator, supporting up to 32 slave resistance-controlled lights to access, and the maximum LAN communication distance is 10 kilometers; the slave resistance-controlled lights communicate with the master resistance-controlled light through dynamic channel allocation, and support automatic reconnection after network disconnection.

[0008] Furthermore, the edge computing module is an edge computing development board equipped with a lightweight insect target detection model, which is deployed after quantization and compression, and uses channel pruning technology to remove redundant convolution kernels.

[0009] Furthermore, both the master resistance-controlled light and the slave resistance-controlled light are integrated with a two-dimensional rotating pan-tilt module with adjustable pitch angle, including a stepper motor drive device and a dynamic steering control strategy.

[0010] Furthermore, the horizontal rotation angle of the stepper motor drive device can be adjusted in the range of 0° to 360°, and the vertical pitch angle can be adjusted in the range of -90° to 90°.

[0011] Furthermore, the dynamic steering control strategy includes: the master resistance-controlled light automatically adjusts the direction of the gimbal according to the cloud command, and sends a set of steering commands to the slave resistance-controlled light through LoRa to form a grating interception array.

[0012] Furthermore, the cloud management platform is equipped with radar data-driven insect peak level determination and dynamic steering algorithm for the pitch angle of the blockage control light, including insect peak level determination and migration behavior parameter solution and decision-making model.

[0013] Furthermore, the determination of insect peak levels and the solution of migration behavior parameters include: real-time analysis of individual target information of migratory pests monitored by insect radar, obtaining their flight altitude, flight speed, and flight direction, and using Fisher's optimal segmentation method to divide and determine the migratory insect peaks per unit time, and output three insect peak levels: high, medium, and low; among them, the high level of insect peaks is that the number of insects per minute is greater than 600, the medium level of insect peaks is that the number of insects per minute is greater than or equal to 300 and less than or equal to 600, and the low level of insect peaks is that the number of insects per minute is less than 300.

[0014] Furthermore, the decision model includes: calculating and outputting the average moving direction θ of the swarm and the pitch angle β at which the swarm is closest to the main control block control light according to the flight altitude, flight speed and flight direction of the multiple targets.

[0015] Furthermore, when the insect peak level is low, the master resistance-controlled light pan / tilt locks the θ direction to keep it the same as the moving direction of the insect swarm, and the master resistance-controlled light and the slave resistance-controlled light turn at the pitch angle β to form an interception sector; when the insect peak level is medium, the master resistance-controlled light performs θ±5° swing scanning, and the slave resistance-controlled light turns at {θ±20°,θ±35°} to form a wide sector; when the insect peak level is high, the master resistance-controlled light switches to the circular scanning mode, and the slave resistance-controlled lights are distributed at equal intervals in a circular array; the species are identified according to edge computing, the wavelength of the insect attractant light is dynamically adjusted, and the optical power of the master resistance-controlled light is adjusted according to the stratification height of the insect swarm, with a maximum irradiation range of 1000 meters.

[0016] Compared with the prior art, the present invention has the following beneficial effects:

[0017] The present invention integrates the flight behavior data of migratory pests monitored by insect radar with multi-spectral light attractant equipment through radar-controlled light fusion, guides the dynamic adjustment of the blockage control strategy, and controls multiple slave blockage control lights to work together through a self-organizing network protocol, thereby reducing costs, greatly increasing the coverage area of ​​light attractants, and saving energy. The power is 800W lower than that of traditional sodium lamps. The edge-cloud collaborative computing architecture is adopted, and the edge computing technology completes pest identification locally. A lightweight insect target detection model is deployed at the lamp end to achieve second-level identification and counting of migratory pests. The cloud regularly pushes the latest model weights to Improve recognition accuracy. In addition, the LED light source dynamically adjusts the spectrum according to the monitoring results of the insect peak types through networking to achieve specific trapping. The barrier control lamp dynamically adjusts the power according to the monitoring results of the insect peak stratification height through networking, and dynamically adjusts the pitch angle and the swing mode according to the monitoring results of the number of insect peaks through networking to achieve automatic control. In short, the present invention realizes accurate monitoring, real-time identification and three-dimensional interception of migratory pests, solves the problems of delayed response, accidental injury to natural enemies, high energy consumption and limited coverage of traditional light trapping equipment, and significantly improves the prevention and control efficiency of migratory pests. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a working schematic diagram of the self-organizing network migratory pest control system based on insect radar and edge computing of the present invention;

[0019] Figure 2 It is a structural schematic diagram of the self-organizing network migratory pest control system based on insect radar and edge computing of the present invention;

[0020] Figure 3 It is a structural schematic diagram of the master resistance-controlled lamp of the present invention;

[0021] Figure 4 This is a workflow diagram of the self-organizing network migratory pest control system based on insect radar and edge computing of the present invention;

[0022] Figure 5 It is a timing diagram of the self-organizing network migratory pest control system based on insect radar and edge computing of the present invention;

[0023] Figure 6 is a plan view schematic diagram of a resistance-controlled lamp of the present invention;

[0024] Figure 7 It is a three-dimensional schematic diagram of the resistance-controlled lamp of the present invention.

[0025] In the figure: 1. impact screen; 2. rotating high-altitude insect-attracting light source; 3. insect-falling funnel; 4. box body; 5. insect-receiving drawer. DETAILED DESCRIPTION

[0026] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and specific embodiments. Embodiment 1

[0027] In the hardware configuration, the master resistance-controlled lamp uses an STM32H743 microcontroller and an edge computing board, equipped with a LoRaSX1276 module and a 4G Cat.1 module; the slave resistance-controlled lamp uses an STM32F407 microcontroller, which integrates LoRa communication and light source driving circuits.

[0028] In the edge computing process, the radar control command is transmitted to the main control resistance control lamp via 4G to control the switch, pitch angle and optical power parameters of the main control resistance control lamp. The lightweight insect target detection model deployed in the edge computing board recognizes the images taken by the camera, outputs the type and number of pests, and calculates the parameters of the heating chamber and the wavelength of the insect attractant lamp.

[0029] In the implementation of self-organizing network communication, the master resistance-controlled light broadcasts the networking beacon, and the slave resistance-controlled light applies for network access through the CSMA / CA mechanism; the master resistance-controlled light assigns a 16-bit short address to the slave resistance-controlled light and issues a time slot synchronization instruction; the slave resistance-controlled light reports status data (voltage, insecticide count) to the master resistance-controlled light every 5 minutes.

[0030] In the collaborative resistance control of master and slave lights, the master resistance-controlled light acts as a data transfer station. On the one hand, it obtains the dynamic execution parameters of cloud commands, and on the other hand, it issues working instructions to adjust the slave resistance-controlled lights through Lora, and collects the information reported by the slave resistance-controlled lights and transmits it back to the cloud.

[0031] In the rotating gimbal control process, radar data is uploaded to the cloud management platform every 10 minutes for parameter recalculation. The cloud transmits the parameters to the edge computing board of the master resistance-controlled light via 4G, triggering the gimbal turning and related operations. The master resistance-controlled light transmits control instructions to the slave resistance-controlled light via Lora, triggering the slave resistance-controlled light gimbal turning and related operations.

[0032] like Figure 1 As shown, the radar monitors the type, layer height and number of insect peaks, and then controls the spectrum, power, pitch angle and swing mode of the resistance control lamp.

[0033] like Figure 2 , Figure 3 , Figure 4 , Figure 5As shown in the figure, the self-organizing network migratory pest control system based on insect radar and edge computing includes a high-resolution insect radar, a master control lamp, a slave control lamp, and a cloud management platform. The high-resolution insect radar monitors the population and flight parameters of migratory pests in the air in real time, and transmits them to the cloud management platform through the network; the master control lamp includes an edge computing module, a multi-spectral trapping module, an infrared insecticide module, and a communication module; the slave control lamp is connected to the master control lamp through a LoRa wireless link, receives light source control instructions, pitch angle parameters and infrared heating parameters, and feeds back working status data; the cloud management platform receives multi-lamp collaborative data through the 4G network, and automatically determines the insect peak according to the real-time monitoring of insect quantity by the insect radar The best insect attraction angle is calculated according to the insect swarm formation, flight direction and speed, and the main light switch, power and pitch angle are remotely and automatically controlled through the 4G communication module to achieve the maximum insect attraction amount; among them, the edge computing module is equipped with a lightweight insect target detection model, the input is the insect body image collected by the visible light camera, and the output is the pest type and density level; the multi-spectral attraction module contains 6 groups of independently modulated LED arrays with a wavelength covering 300-650nm, and dynamically combines the light-emitting bands and power ups and downs according to the recognition results; the infrared insecticide module adopts an infrared heating chamber to adjust the heating temperature and heating time according to the pest type, and adapt to the thresholds of pests of different sizes; the communication module integrates the LoRa SX1276 chip and the 4G Cat.1 module to receive instructions sent by the insect radar, and realize low-power networking with the slave barrier control lamp and cloud data backhaul. The LoRa self-organizing network protocol adopts a star topology structure. The master resistance-controlled light is the network coordinator, which supports up to 32 slave resistance-controlled lights to access, and the maximum LAN communication distance is 10 kilometers. The slave resistance-controlled light communicates with the master resistance-controlled light through dynamic channel allocation, and supports automatic reconnection after network disconnection. The edge computing module is an edge computing development board, equipped with a lightweight insect target detection model, which is deployed after quantization compression, and uses channel pruning technology to remove redundant convolution kernels. Both the master resistance-controlled light and the slave resistance-controlled light are integrated with a two-dimensional rotating pan-tilt module with adjustable pitch angle, including a stepper motor drive device and a dynamic steering control strategy. The horizontal rotation angle of the stepper motor drive device can be adjusted from 0° to 360°, and the vertical pitch angle can be adjusted from -90° to 90°. The dynamic steering control strategy includes: the master resistance-controlled light automatically adjusts the pan-tilt direction according to the cloud command, and sends a steering command set to the slave resistance-controlled light through LoRa to form a grating interception array. The cloud management platform is equipped with radar data-driven insect peak level determination and dynamic steering algorithm for the pitch angle of the blockage control light, including insect peak level determination and migration behavior parameter calculation as well as decision-making models.The determination of insect peak levels and the calculation of migration behavior parameters include: real-time analysis of individual target information of migratory pests monitored by insect radar, obtaining their flight altitude, flight speed, and flight direction, using Fisher's optimal segmentation method to divide and determine the migratory insect peaks in unit time, and output three insect peak levels: high, medium, and low; among them, the high level of insect peaks is more than 600 insects per minute, the medium level of insect peaks is more than or equal to 300 and less than or equal to 600 insects per minute, and the low level of insect peaks is less than 300 insects per minute. The decision model includes: according to the flight altitude, flight speed, and flight direction of multiple targets, the average moving direction θ of the output group and the pitch angle β of the insect group closest to the main control block control light are calculated. When the insect peak level is low, the master resistance-controlled light pan / tilt locks the θ direction to keep it the same as the movement direction of the insect swarm, and the master resistance-controlled light and the slave resistance-controlled light turn at the pitch angle β to form an interception sector; when the insect peak level is medium, the master resistance-controlled light performs θ±5° swing scanning, and the slave resistance-controlled light turns at {θ±20°,θ±35°} to form a wide sector; when the insect peak level is high, the master resistance-controlled light switches to circular scanning mode, and the slave resistance-controlled lights are evenly spaced in a circular array; the species are identified according to edge computing, the wavelength of the insect attractant is dynamically adjusted, and the optical power of the master resistance-controlled light is adjusted according to the height of the insect swarm stratification, with a maximum irradiation range of 1000 meters.

[0034] like Figure 6 and Figure 7 As shown, the impedance lamp is composed of an impact screen 1, a rotating high-altitude insect-attracting light source 2, an insect-falling funnel 3, a box body 4, and an insect-receiving drawer 5.

[0035] The resistance-controlled light adopts a foldable bracket and a waterproof shell (IP67), with a total weight of less than or equal to 5kg. In the rapid deployment mode, the inclination angle is automatically calibrated after being inserted into the ground, and the networking is completed within 10 minutes. In the gimbal mechanical design, a waterproof stepper motor (model 28BYJ-48) is used with a harmonic reducer, and the torque is greater than or equal to 0.25N·m. The rotating mechanism has a built-in Hall sensor to achieve a positioning accuracy of ±0.5°. In the dynamic balance design, a counterweight block is configured at the bottom of the lamp body to ensure stable operation under level 6 wind conditions. The resistance-controlled light is equipped with four universal wheels for easy transportation and mobile deployment.

[0036] In summary, radar-edge computing collaboration achieves accurate identification of migratory pests and generation of dynamic interception strategies through real-time fusion of insect radar data and lightweight insect target detection models; in the LoRa+4G hybrid network, low-power LoRa communication is used between master and slave lamps, and 4G network is used for cloud interaction, taking into account real-time performance, cost and wide-area coverage of control range; radar-guided intelligent steering interception analyzes radar data in real time, and the cloud platform drives the blocking control lamp array to form an adaptive steering "light barrier"; real-time monitoring-identification-interception closed-loop control of migratory pests is realized, which is suitable for three-dimensional prevention and control of migratory pests in fields, forestry and grasslands.

[0037] The above-mentioned specific embodiments are only several preferred embodiments of the present invention. Based on the technical solutions of the present invention and the relevant inspirations of the above-mentioned embodiments, those skilled in the art can make various alternative improvements and combinations to the above-mentioned specific embodiments.

Claims

1. A self-organizing network migratory pest control system based on insect radar and edge computing, including high-resolution insect radar, master control lamp, slave control lamp, and cloud management platform, characterized by: The high-resolution insect radar monitors the population and flight parameters of migrating pests in the air in real time, and transmits them to the cloud management platform through the network; the master control light includes an edge computing module, a multi-spectral trapping module, an infrared insecticide module, and a communication module; the slave control light is connected to the master control light through a LoRa wireless link, receives light source control instructions, pitch angle parameters and infrared heating parameters, and feeds back working status data; the cloud management platform receives multi-light coordination data through the 4G network, automatically determines the insect peak level based on the real-time monitoring of the insect quantity by the insect radar, calculates the best insect attraction angle based on the insect swarm formation, flight direction and speed, and remotely and automatically controls the main light switch, power and pitch angle through the 4G communication module to achieve the maximum insect attraction; Among them, the edge computing module is equipped with a lightweight insect target detection model, the input is the insect body image collected by the visible light camera, and the output is the pest type and density level; the edge computing module is an edge computing development board, equipped with a lightweight insect target detection model, which is deployed after quantization compression, and uses channel pruning technology to remove redundant convolution kernels; the multi-spectral trapping module contains 6 groups of independently modulated LED arrays, with a wavelength covering 300-650nm, which adjusts the spectrum to adapt to the preferences of different pests, and dynamically combines the luminous bands and power ups and downs according to the recognition results; the infrared insecticide module uses an infrared heating chamber to adjust the heating temperature and heating time according to the pest type, and adapt to the thresholds of pests of different sizes; the communication module integrates the LoRa SX1276 chip and 4G Cat.1 module, which can receive commands sent by insect radar, realize low-power networking with slave resistance-controlled lights and cloud data transmission; LoRa self-organizing network protocol adopts star topology, the master resistance-controlled light is the network coordinator, and supports up to 32 slave resistance-controlled lights to access, and the maximum LAN communication distance is 10 kilometers; the slave resistance-controlled lights communicate with the master resistance-controlled light through dynamic channel allocation, and support automatic reconnection after network disconnection.

2. The self-organizing network migratory pest control system based on insect radar and edge computing according to claim 1 is characterized in that: Both the master resistance-controlled light and the slave resistance-controlled light are integrated with a two-dimensional rotating pan-tilt module with adjustable pitch angle, including a stepper motor drive device and a dynamic steering control strategy.

3. The self-organizing network migratory pest control system based on insect radar and edge computing according to claim 2 is characterized in that: The horizontal rotation angle of the stepper motor drive device can be adjusted from 0° to 360°, and the vertical pitch angle can be adjusted from -90° to 90°.

4. The self-organizing network migratory pest control system based on insect radar and edge computing according to claim 2 is characterized in that: The dynamic steering control strategy includes: the master resistance-controlled light automatically adjusts the gimbal direction according to the cloud command, and sends a set of steering commands to the slave resistance-controlled light through LoRa to form a grating interception array.

5. The self-organizing network migratory pest control system based on insect radar and edge computing according to claim 1 is characterized in that: The cloud management platform is equipped with radar data-driven insect peak level determination and dynamic steering algorithm for the pitch angle of the blockage control light, including insect peak level determination and migration behavior parameter calculation as well as decision-making models.

6. The self-organizing network migratory pest control system based on insect radar and edge computing according to claim 5 is characterized in that: The determination of insect peak levels and the calculation of migration behavior parameters include: real-time analysis of individual target information of migratory pests monitored by insect radar, obtaining their flight altitude, flight speed, and flight direction, and using Fisher's optimal segmentation method to divide and determine the migratory insect peaks per unit time, and output three insect peak levels: high, medium, and low; among them, the high level of insect peaks is that the number of insects per minute is greater than 600, the medium level of insect peaks is that the number of insects per minute is greater than or equal to 300 and less than or equal to 600, and the low level of insect peaks is that the number of insects per minute is less than 300.

7. The self-organizing network migratory pest control system based on insect radar and edge computing according to claim 6, characterized in that: The decision-making model includes: calculating and outputting the average moving direction θ of the swarm and the pitch angle β at which the swarm is closest to the main control block light according to the flight altitude, flight speed and flight direction of multiple targets.

8. The self-organizing network migratory pest control system based on insect radar and edge computing according to claim 7 is characterized in that: When the insect peak level is low, the master resistance-controlled light pan / tilt locks the θ direction to keep it the same as the movement direction of the insect swarm, and the master resistance-controlled light and the slave resistance-controlled light turn at the pitch angle β to form an interception sector; when the insect peak level is medium, the master resistance-controlled light performs θ±5° swing scanning, and the slave resistance-controlled light turns at {θ±20°,θ±35°} to form a wide sector; when the insect peak level is high, the master resistance-controlled light switches to circular scanning mode, and the slave resistance-controlled lights are evenly spaced in a circular array; the species are identified according to edge computing, the wavelength of the insect attractant is dynamically adjusted, and the optical power of the master resistance-controlled light is adjusted according to the height of the insect swarm stratification, with a maximum irradiation range of 1000 meters.

Images