Agricultural pest monitoring system and method

Abstract

The present invention provides an agricultural pest monitoring system and method, wherein a plurality of monitoring points are set in a target area, each monitoring point includes a data collection device for real-time collection of pest activity data; a data processing subsystem receives and processes the pest activity data to obtain pest monitoring result data of each monitoring point; and the pest monitoring result data of the monitoring point is processed to obtain pest monitoring result data of the target area; a remote database stores pest activity data and pest monitoring result data. The system is real-time, efficient, and low-cost, and pest detection can be performed in real time over a long period of time, and the monitoring accuracy and work autonomy are improved.

Description

Technical Field

[0001] The present invention relates to the technical field related to remote automated measurement, and in particular to an agricultural pest monitoring system and method. Background Art

[0002] Agriculture is the foundation of a country, and agricultural security is directly related to the security of national strategies. my country has been committed to improving the level of agricultural informatization and technology through various means, thereby improving agricultural production efficiency, increasing product supply and output to society, and ultimately meeting social needs. However, the production sites involved in various branches of agriculture cover a wide area, have long cycles, and have large differences in conditions. How to effectively monitor agricultural pests and diseases has always been an important pain point and difficulty in further improving agricultural production efficiency and balancing the green development concept.

[0003] Existing agricultural pest monitoring technologies include satellite spectral remote sensing technology, drone remote sensing technology, Internet of Things machine vision technology, etc. Among them, satellite spectral remote sensing technology is suitable for a larger coverage range and often has limited resolution; drone remote sensing technology is affected by flight costs and flight conditions, and cannot conduct long-term real-time pest monitoring in agricultural production areas; machine vision technology based on the agricultural Internet of Things can, to a certain extent, monitor the observation area in real time over a long period of time, but because it only relies on the image data of the observed area for analysis and calculation, and the pest images need to be marked in advance in order to train the machine vision learning model, the monitoring accuracy and work autonomy are limited. Summary of the invention

[0004] Based on the above problems, the present invention provides an agricultural pest monitoring system and method, aiming to solve technical problems in the prior art such as the inability to perform long-term and real-time pest detection, limited monitoring accuracy and working autonomy, etc.

[0005] An agricultural pest monitoring system, comprising:

[0006] A number of monitoring points are set up in the target area, each monitoring point includes a data collection device for collecting pest activity data in real time;

[0007] The data acquisition device, a data processing subsystem and a remote database are connected to each other through a network transmission device;

[0008] The data processing subsystem is used to:

[0009] Receiving and processing pest activity data to obtain pest monitoring result data of each monitoring point; and processing the pest monitoring result data of the monitoring point to obtain pest monitoring result data of the target area;

[0010] Remote database for storing pest activity data and pest monitoring result data.

[0011] Furthermore, the data acquisition device comprises:

[0012] substrate;

[0013] A pressure sensor, disposed on at least one surface of the substrate, for acquiring pressure data of the surface of the substrate;

[0014] An image sensor is fixed to the substrate via a bracket and is used to obtain image data of the surface of the substrate;

[0015] Pest activity data include pressure data and image data;

[0016] The data processing subsystem is used to:

[0017] receiving pressure data of the monitoring point and forming first time series data according to the time stamp, receiving image data and forming second time series data according to the time stamp; and

[0018] Processing the first time series data to extract first feature data; and

[0019] Processing the second time series data to extract second feature data; and

[0020] For each monitoring point, the first feature data and the second feature data are combined in the order of timestamps to form a time series feature set; and

[0021] Using the time series feature set as input, clustering calculation is performed using a clustering algorithm model to obtain pest monitoring result data of the monitoring point; and

[0022] According to the layout principle and coverage area of ​​monitoring points in the target area, the pest monitoring result data of the monitoring points calculated by the clustering calculation model are processed by the systematic sampling calculation method to obtain the pest monitoring result data of the target area.

[0023] Furthermore, the data processing subsystem is also used for:

[0024] Call the time series feature set within the preset time period;

[0025] Taking the time series data set within the preset time period as input, the trained classification calculation model is used to perform classification calculation to obtain the pest monitoring result data of the monitoring point within the preset time period; and

[0026] According to the layout principles and coverage areas of monitoring points in the target area, the pest monitoring result data of the monitoring points calculated by the classification calculation model are processed by the systematic sampling calculation method to obtain the pest monitoring result data of the target area.

[0027] Furthermore, the pressure data includes the acquisition timestamp, the point of application, the direction of application, and the magnitude of application;

[0028] The first characteristic data includes the pest's moving speed, moving direction, each moving stop time, and the change value of the force direction;

[0029] The second feature data includes the movement trajectory of the pests.

[0030] Furthermore, the data acquisition device also includes:

[0031] An auxiliary acquisition module, used for acquiring auxiliary data of the surface of the substrate;

[0032] The data processing center is also used to:

[0033] receiving auxiliary data of the monitoring point and forming third time series data according to the timestamp;

[0034] Processing the third time series data to extract third feature data;

[0035] For each monitoring point, the first feature data, the second feature data and the third feature data are combined in the order of timestamps to form a time series feature set.

[0036] Furthermore, the pest monitoring result data includes the types of pests, the number and distribution of each pest;

[0037] The pest monitoring result data calculated by the classification calculation model also includes the degree of pest activity.

[0038] Furthermore, the auxiliary sensor is at least one of the following sensors: a sound sensor, an infrared sensor, and a vibration sensor.

[0039] A method for monitoring agricultural pests, using the aforementioned agricultural pest monitoring system, sets a number of monitoring points in a target area, including:

[0040] Step A1, real-time collection of pest activity data at each monitoring point;

[0041] Step A2, processing the pest activity data to obtain pest monitoring result data of the monitoring point, and obtaining pest monitoring result data of the target area according to the pest monitoring result data;

[0042] Step A3, storing pest activity data and pest monitoring result data.

[0043] Furthermore, the pest activity data includes pressure data and image data, and step A2 includes:

[0044] Step A201, receiving pressure data of a monitoring point and forming first time series data according to a timestamp, receiving image data and forming second time series data according to a timestamp;

[0045] Step A202, processing the first time series data to extract first characteristic data, and processing the second time series data to extract second characteristic data;

[0046] Step A203: for each monitoring point, the first feature data and the second feature data are combined in the order of timestamps to form a time series feature set;

[0047] Step A204, using the time series feature set as input, using a clustering algorithm model to perform clustering calculations to obtain pest monitoring result data of the monitoring point;

[0048] Step A205, based on the layout principle and coverage area of ​​the monitoring points in the target area, the pest monitoring result data of the monitoring points calculated by the clustering calculation model are processed by a systematic sampling calculation method to obtain the pest monitoring result data of the target area.

[0049] Furthermore, step A2 also includes:

[0050] Step A206, sending the time series feature set to a remote database for storage;

[0051] Step A207, calling a time series feature set within a preset time period from a remote database;

[0052] Step A208, taking the time series data set within the preset time period as input, using the trained classification calculation model to perform classification calculation, and obtaining the pest monitoring result data of the monitoring point within the preset time period;

[0053] Step A209, based on the layout principle and coverage area of ​​the monitoring points in the target area, the pest monitoring result data of the monitoring points calculated by the classification calculation model are processed by a systematic sampling calculation method to obtain the pest monitoring result data of the target area.

[0054] The beneficial technical effect of the present invention is that: a number of monitoring points are set up in the target area, pest activity data are collected, and the data are processed to obtain monitoring result data of the monitoring points and the target area. It is real-time, efficient, and low-cost, and pest detection can be performed in real time over a long period of time, and the monitoring accuracy and work autonomy are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1-2 A schematic diagram of a module of an agricultural pest monitoring system of the present invention;

[0056] Figure 3A schematic diagram of the structure of a data acquisition device for an agricultural pest monitoring system according to the present invention;

[0057] Figure 4-5 A schematic diagram of the arrangement of pressure sensors of an agricultural pest monitoring system of the present invention;

[0058] Figure 6-8 The present invention is a flowchart of the steps of an agricultural pest monitoring method. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0061] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0062] See also Figure 1 The present invention provides an agricultural pest monitoring system, comprising:

[0063] A plurality of monitoring points are set up in the target area, each monitoring point includes a data collection device (1) for collecting pest activity data in real time;

[0064] The data acquisition device (1), the data processing subsystem (2) and the remote database (4) are connected to each other through a network transmission device (3);

[0065] The data processing subsystem (2) is used to:

[0066] Receiving and processing pest activity data to obtain pest monitoring result data of each monitoring point; and processing the pest monitoring result data of the monitoring point to obtain pest monitoring result data of the target area;

[0067] The remote database (4) is used to store pest activity data and pest monitoring result data.

[0068] See also Figure 2 , further, the data acquisition device (1) comprises:

[0069] substrate;

[0070] A pressure sensor (102), disposed on at least one surface of the substrate, for acquiring pressure data of the surface of the substrate;

[0071] An image sensor (103) is fixed to the substrate via a bracket and is used to obtain image data of the surface of the substrate;

[0072] Pest activity data include pressure data and image data;

[0073] The data processing subsystem (2) is used to:

[0074] receiving pressure data of the monitoring point and forming first time series data according to the time stamp, receiving image data and forming second time series data according to the time stamp; and

[0075] Processing the first time series data to extract first feature data; and

[0076] Processing the second time series data to extract second feature data; and

[0077] For each monitoring point, the first feature data and the second feature data are combined in the order of timestamps to form a time series feature set; and

[0078] Using the time series feature set as input, clustering calculation is performed using a clustering algorithm model to obtain pest monitoring result data of the monitoring point; and

[0079] According to the layout principle and coverage area of ​​monitoring points in the target area, the pest monitoring result data of the monitoring points calculated by the clustering calculation model are processed by the systematic sampling calculation method to obtain the pest monitoring result data of the target area.

[0080] Specifically, before clustering calculation, the time series feature set is preprocessed, and the preprocessing includes feature selection, dislocation transformation, and time series discretization. Feature selection can remove irrelevant features and redundant features to achieve feature dimensionality reduction. Since there is a certain correlation between time series data in time, processing the data in the data set alone will lose this correlation, but too much involvement will cause the data dimension to increase sharply, resulting in excessive computational complexity. Preferably, the time series data set can be transformed into a new data set that carries the correlation of the time series data without increasing the dimension too much by dislocation transformation.

[0081] Repair missing values ​​in the preprocessed time series feature set. In the process of dislocation transformation, data loss may occur. Optionally, Gibbs sampling method is used to repair the missing class data. The process of repairing missing data is an iteration, starting from the first missing data, repairing one at a time, and gradually filling all the missing data. Finally, a new time series feature set is obtained.

[0082] The clustering algorithm model is pre-trained, and data is collected in advance as sample data. The sample data is divided into a training set, a test set, and a validation set to calculate the clustering algorithm model.

[0083] Different clustering algorithms can be selected for different application scenarios. Specifically, K-means clustering algorithm, density clustering, model clustering and other algorithm models can be used. Preferably, a small data set can be tested to select a suitable clustering algorithm model and determine the preferred parameter values ​​of the clustering algorithm model.

[0084] The components of the monitoring system of the present invention include a plurality of data acquisition devices (1) of the target area, a data processing subsystem (2), a remote database (4) and a network transmission device (3).

[0085] In some embodiments, the components in the application scenario can be interconnected through a network transmission device (3), including wireless links, wired links, or hybrid links. The monitoring points can communicate and interact with each other through the network transmission device (3), and the monitoring points can be connected to a remote database through the network transmission device (3) for data transmission. The data processing subsystem (2) is connected to the monitoring point through the network transmission device (3) to obtain and query real-time data. The data processing subsystem (2) is connected to a remote database (4) through the network transmission device (3) for access.

[0086] The monitoring points are deployed and configured according to the monitoring accuracy requirements of the target area. They are mainly used to collect pest activity data, including the movement trajectory of the pests, the direction and magnitude of the force applied by the feet during the movement of the pests, and pest image data.

[0087] Specifically, the data processing subsystem (2) includes a data processing center and local computing devices set up at each monitoring point, and all local computing devices communicate with the data processing center through a network transmission device (3);

[0088] As an embodiment, not only a data collection device (1) is arranged at the monitoring point, but also a local computing device is configured. The local computing device collects the pest activity data of the data collection device (1) at the monitoring point, performs clustering calculation of insect species, and obtains the pest monitoring result data of the local monitoring point through clustering calculation.

[0089] As another embodiment, each monitoring point transmits data through a network transmission device (3), and a local computing device can obtain the pest activity data of other monitoring points and perform clustering calculations to obtain the pest monitoring result data of each monitoring point, and then process the data based on a systematic sampling calculation method to obtain the pest monitoring result data of the target area.

[0090] As a preferred embodiment, the local computing device will collect pest activity data and upload it to the data processing center, or the data processing center can select the pest activity data of one or more specific monitoring points for cluster calculation to obtain the pest monitoring result data of each monitoring point, and then obtain the pest monitoring result data of the target area based on the systematic sampling calculation method.

[0091] That is to say, the pest monitoring result data of the monitoring point and the pest monitoring result data of the target area can be processed on a local computing device or in a data processing center.

[0092] Specifically, the deployment of monitoring points can be determined according to the pest monitoring accuracy required for the target area. At the same time, the deployment density of monitoring points may also be affected by factors such as the pest types, pest distribution diversity, crop types, and crop density in the target area.

[0093] Specifically, the choice of communication module for monitoring points can be selected according to the geographical environment conditions of the target area. In the absence of infrastructure support and in harsh environmental conditions, satellite communication modules can be used for network communication; in the presence of infrastructure support, 4G / 5G communication modules can be used to access the network for two-way communication; in application scenarios with more convenient conditions, wireless LAN or Ethernet interface cards can be used to access the network to provide network services with higher data transmission rates and lower communication delays.

[0094] Specifically, the network transmission equipment (3) includes two parts: a transmission link and a transmission device. The transmission link may include a short-distance wireless transmission link, a wired link, a satellite communication link, etc.; the network transmission equipment (3) includes a network interface card, a transmission antenna, a router, a switch, and other hardware devices related to data transmission and reception.

[0095] Specifically, the remote database (4) may be a cloud database or a local database, which has enough space to accommodate the generated data and can meet the concurrent access requirements of a large number of monitoring points and all devices in the data processing center.

[0096] Specifically, the data processing center (2) may include mobile terminal devices, processing node devices, and a data visualization screen. The mobile terminal device is often connected to the system through a wireless network, and can access any monitoring point to view real-time pest activity data and pest monitoring result data. It can also access historical data in a remote database or real-time data sent by a monitoring node through a network device. The processing node device can be a computing device with strong computing power, a local distributed computing cluster that can simultaneously classify and calculate a large number of real-time or non-real-time data models, or a cloud host that performs distributed computing operations on pest activity data through remote distributed computing capabilities. The data visualization screen is mainly used to present the real-time monitoring point network status, the visualization of pest activity data, and the visualization of the calculation results of the target area pest situation.

[0097] It should be noted that the description of the system and its components, especially the network equipment part, described in the figure is only an example for the convenience of description, and this specification cannot be limited to the scope of application scenarios of the described embodiments. It should be understood that for engineers and technicians in this field, after understanding the principles of the system, they can arbitrarily combine the various components without deviating from the system and its working principle, or form a subsystem to connect with other components. For example, changes in the shape of the micro-pressure sensor node substrate, combining monitoring points of multiple different types of sensor nodes, flexible changes in monitoring point modules, multiple remote databases in the system, a multi-level processing center in the system, and the merger of the processing center and the remote database. Such deformations are all within the scope of protection of this specification.

[0098] See also Figure 3 Specifically, the data acquisition device (1) at the monitoring point includes a rectangular (circular) substrate on which an array of micro pressure sensors (102) is laid, and an image sensor (103);

[0099] The data acquisition device (1) further comprises:

[0100] A universal arm (101) connecting the substrate and the image sensor (103);

[0101] The monitoring points also include local computing equipment, communication modules, antennas, power supplies, etc.

[0102] Specifically, the substrate serves as a sampling space for pest activities in the target area, and includes an upper surface and a lower surface, and at least one of the surfaces is provided with a groove for laying a micro-pressure sensor.

[0103] The pressure sensor (102) is specifically a micro pressure sensor array, which is used to sense and capture the behavioral characteristics of pests entering the sampling space at the monitoring point. The laying points and laying density of the micro pressure sensors are determined according to specific needs. Grooves are opened in advance on the substrate to be laid to fix the micro pressure sensor nodes and connect data lines.

[0104] Specifically, the micro-pressure sensor array is configured to collect the force point, force direction and pressure value of each foot of multi-legged pests crawling on the substrate. This sensor array can capture the corresponding data and output digital or analog sensor signals. The sensor array can be configured to periodically sense the sampling space and output the corresponding data, or it can be configured as an event-triggered working mode, performing perception operations and outputting pressure data only when an event of interest occurs.

[0105] Specifically, the image sensor (103) uses a universal arm (101) that can be adjusted in any direction as a bracket, one end of which is connected to the substrate of the monitoring point, and the other end supports the image sensor (103) to cover the entire rectangular (circular) monitoring area substrate in the form of a probe, and is configured to capture images of pest movement in the monitoring area, or is configured to dynamically track and collect trajectory data of pest movement. In some embodiments, the image sensor has no fixed parameter restrictions, but needs to be selected according to the specific application scenario, including but not limited to the following parameters, such as image resolution, night vision function, infrared function, storage capacity, size, power supply method, communication interface, data format, etc.

[0106] Specifically, it includes one or more local computing devices with storage functions, which are configured to receive the output data of the micro pressure sensor array and the image sensor, perform clustering calculations, and output the results. In addition to performing local calculations, the local computing device will also periodically transmit the collected pest activity data to a remote database (4); it may also receive and respond to data query requests initiated by a data processing center or other local computing devices.

[0107] Specifically, it includes one or more communication modules including antennas, which are configured to transmit the output results of the local computing device to other devices or network nodes, and can also receive control commands such as data query requests sent by other nodes or devices.

[0108] The pest activity data collected by the monitoring point can be configured as real-time feedback, periodic feedback, or custom rule feedback as needed. If configured as real-time feedback, the monitoring point will send the collected pressure data and image data to the remote database for storage in real time; if configured as periodic feedback, the monitoring point will start a feedback operation at regular intervals to send the collected pest activity data back to the remote database; in addition, a threshold for the amount of data to be saved can also be pre-configured. When the amount of data saved locally at the monitoring node reaches the threshold, the feedback operation is started.

[0109] Specifically, Figure 4 A schematic diagram of laying out a micro pressure sensor array on a rectangular substrate is given, and the intersection of line segments parallel to the X-axis and the Y-axis is selected as the layout point position of the micro pressure sensor, wherein the number of micro pressure sensor nodes to be laid out is determined by the spacing distance between the line segments parallel to the X-axis and the Y-axis. According to the monitoring accuracy requirements of a specific embodiment, different spacings can be set to calculate the number of micro pressure sensor nodes that need to be laid out. Figure 5 A schematic diagram of laying out a micro-pressure sensor array on a circular substrate is provided, and several intersection points of diameters of the concentric circles on the outer edge of the substrate and the uniformly divided substrate are selected as the locations of the micro-pressure sensor layout points, wherein the number of micro-pressure sensor nodes to be laid out is determined by the difference in radius of the concentric circles and the number of diameters of the uniformly divided substrate. According to the monitoring accuracy requirements of the embodiment, the number of micro-pressure sensor layout points can be calculated, and the specific micro-pressure sensor layout points can be determined by weighing the difference in radius of the concentric circles and the number of diameters of the uniformly divided substrate. The pressure data of the monitored pests moving on (under) the substrate is read in real time from the local micro-pressure sensor array at the monitoring point and saved in a local storage device. Specifically, clustering calculations can be performed on the pest activity data collected at the monitoring point based on a semi-supervised machine learning algorithm.

[0110] Specifically, clustering calculations can be performed on the collected pest activity data based on an unsupervised machine learning algorithm.

[0111] Furthermore, the data processing subsystem (2) is also used for:

[0112] Call the time series feature set within the preset time period;

[0113] Taking the time series data set within the preset time period as input, the trained classification calculation model is used to perform classification calculation to obtain the pest monitoring result data of the monitoring point within the preset time period; and

[0114] According to the layout principles and coverage areas of monitoring points in the target area, the pest monitoring result data of the monitoring points calculated by the classification calculation model are processed by the systematic sampling calculation method to obtain the pest monitoring result data of the target area.

[0115] Specifically, the time series feature set is sent to the remote database (4), and the data processing subsystem can call the time series feature set within a certain time interval from the remote database (4);

[0116] Specifically, the time series feature set sent by the remote database within a preset time period is preprocessed, and the preprocessing includes feature selection, data discretization, dislocation transformation, etc., to obtain a new time series feature set.

[0117] The classification calculation model is pre-trained, and the classification calculation model is an RNN or LSTM model.

[0118] The classification calculation model can process pest activity data at any monitoring point of interest and in any time period.

[0119] As an embodiment, not only a data collection device (1) is arranged at the monitoring point, but also a local computing device is configured, and the local computing device collects the pest activity data of the data collection device (1) at the monitoring point; the local computing device regularly uploads the pest activity data to the remote database (4).

[0120] The local computing device performs classification calculations on the insect species, and obtains the pest monitoring result data of the local monitoring point through the classification calculations.

[0121] As another embodiment, each monitoring point transmits data through a network transmission device (3), and a local computing device can obtain the pest activity data of other monitoring points and perform classification calculations to obtain the pest monitoring result data of each monitoring point, and then process it based on a systematic sampling calculation method to obtain the pest monitoring result data of the target area.

[0122] As a preferred embodiment, the local computing device will collect pest activity data and upload it to the data processing center, or the data processing center can select the pest activity data of one or more specific monitoring points for classification calculation, obtain the pest monitoring result data of each monitoring point, and then process it based on the systematic sampling calculation method to obtain the pest monitoring result data of the target area. The local computing device regularly uploads the pest activity data and time series feature set to the remote database (4). The data processing center can call the time series feature set within a preset time period from the remote database (4) for classification calculation.

[0123] That is to say, the acquisition of the pest monitoring result data of the monitoring point and the pest monitoring result data of the target area can be performed on a local computing device or in a data processing center.

[0124] The pest activity data pre-collected by the data collection device (1) is periodically stored in the remote database (4). Subsequently, the time series data is manually marked as sample data, and a classification calculation model for pest classification is established and trained, so that the pest activity data of any selected time period can be classified and calculated, and the pest distribution in the target detection area can be estimated. In particular, the processing center 130 can select the collected data of a specific monitoring point for classification calculation and pest distribution estimation, and can also select the collected data of multiple monitoring points in a target area for classification calculation and pest distribution estimation.

[0125] Furthermore, the pressure data includes the acquisition timestamp, the point of application, the direction of application, and the magnitude of application;

[0126] The first characteristic data includes the pest's moving speed, moving direction, each moving stop time, and the change value of the force direction;

[0127] The second feature data includes the movement trajectory of the pests.

[0128] Furthermore, the data acquisition device (1) further comprises:

[0129] An auxiliary acquisition module (104), used for acquiring auxiliary data of the surface of the substrate;

[0130] The data processing center (2) is also used for:

[0131] receiving auxiliary data of the monitoring point and forming third time series data according to the timestamp;

[0132] Processing the third time series data to extract third feature data;

[0133] For each monitoring point, the first feature data, the second feature data and the third feature data are combined in the order of timestamps to form a time series feature set.

[0134] Furthermore, the pest monitoring result data includes the types of pests, the number and distribution of each pest;

[0135] The pest monitoring result data calculated by the classification calculation model also includes the degree of pest activity.

[0136] Furthermore, the auxiliary sensor (104) is at least one of the following sensors: a sound sensor, an infrared sensor, and a vibration sensor.

[0137] See also Figure 6 The present invention also provides an agricultural pest monitoring method, using the aforementioned agricultural pest monitoring system to set a number of monitoring points in the target area, including:

[0138] Step A1, real-time collection of pest activity data at each monitoring point;

[0139] Step A2, processing the pest activity data to obtain pest monitoring result data of the monitoring point, and obtaining pest monitoring result data of the target area according to the pest monitoring result data;

[0140] Step A3, storing pest activity data and pest monitoring result data.

[0141] See also Figure 7 , further, the pest activity data includes pressure data and image data, and step A2 includes:

[0142] Step A201, receiving pressure data of a monitoring point and forming first time series data according to a timestamp, receiving image data and forming second time series data according to a timestamp;

[0143] Step A202, processing the first time series data to extract first characteristic data, and processing the second time series data to extract second characteristic data;

[0144] Step A203: for each monitoring point, the first feature data and the second feature data are combined in the order of timestamps to form a time series feature set;

[0145] Step A204, using the time series feature set as input, using a clustering algorithm model to perform clustering calculations to obtain pest monitoring result data of the monitoring point;

[0146] Step A205, based on the layout principle and coverage area of ​​the monitoring points in the target area, the pest monitoring result data of the monitoring points calculated by the clustering calculation model are processed by a systematic sampling calculation method to obtain the pest monitoring result data of the target area.

[0147] See also Figure 8 , further, step A2 also includes:

[0148] Step A206, sending the time series feature set to the remote database (4) for storage;

[0149] Step A207, calling the time series feature set within a preset time period from the remote database (4);

[0150] Step A208, taking the time series data set within the preset time period as input, using the trained classification calculation model to perform classification calculation, and obtaining the pest monitoring result data of the monitoring point within the preset time period;

[0151] Step A209, based on the layout principle and coverage area of ​​the monitoring points in the target area, the pest monitoring result data of the monitoring points calculated by the classification calculation model are processed by a systematic sampling calculation method to obtain the pest monitoring result data of the target area.

[0152] The above are only preferred embodiments of the present invention, and are not intended to limit the implementation methods and protection scope of the present invention. Those skilled in the art should be aware that all solutions obtained by equivalent substitutions and obvious changes made using the description and illustrations of the present invention should be included in the protection scope of the present invention.

Claims

1. An agricultural pest monitoring system, characterized in that: include: A plurality of monitoring points are set up in the target area, each of the monitoring points comprises a data collection device for collecting pest activity data in real time; The data acquisition device, a data processing subsystem and a remote database are connected to each other through a network transmission device; The data processing subsystem is used for: Receiving and processing the pest activity data to obtain pest monitoring result data of each monitoring point; and processing the pest monitoring result data of the monitoring point to obtain the pest monitoring result data of the target area; The remote database is used to store the pest activity data and the pest monitoring result data; The data acquisition device comprises: substrate; A pressure sensor, disposed on at least one surface of the substrate, for acquiring pressure data of the surface of the substrate; An image sensor, fixed to the substrate via a bracket, for acquiring image data of the surface of the substrate; The pest activity data includes the pressure data and the image data; The data processing subsystem is used for: receiving the pressure data of the monitoring point and forming first time series data according to the timestamp, receiving the image data and forming second time series data according to the timestamp; and Processing the first time series data to extract first feature data; and Processing the second time series data to extract second feature data; and For each of the monitoring points, merging the first feature data and the second feature data in the order of timestamps to form a time series feature set; and Taking the time series feature set as input, a clustering algorithm model is used to perform clustering calculation to obtain the pest monitoring result data of the monitoring point; and According to the layout principle and coverage area of ​​the monitoring points in the target area, the pest monitoring result data of the monitoring points calculated by the clustering algorithm model are processed by a systematic sampling calculation method to obtain the pest monitoring result data of the target area; The pressure data includes acquisition timestamp, force point, force direction, and force magnitude; The first characteristic data includes the pest's moving speed, moving direction, each movement stop time, and the change value of the force direction; The second characteristic data includes the movement trajectory of the pests.

2. An agricultural pest monitoring system as claimed in claim 1, characterized in that: The data processing subsystem is also used for: Calling the time series feature set within a preset time period; Taking the time series data set within the preset time period as input, a trained classification calculation model is used to perform classification calculation to obtain the pest monitoring result data of the monitoring point within the preset time period; as well as According to the layout principle and coverage area of ​​the monitoring points in the target area, the pest monitoring result data of the monitoring points calculated by the classification calculation model are processed by a systematic sampling calculation method to obtain the pest monitoring result data of the target area.

3. An agricultural pest monitoring system as claimed in claim 1, characterized in that: The data acquisition device also includes: An auxiliary acquisition module, used for acquiring auxiliary data of the surface of the substrate; The data processing subsystem is also used for: receiving the auxiliary data of the monitoring point and forming third time series data according to the timestamp; Processing the third time series data to extract third feature data; For each of the monitoring points, the first feature data, the second feature data and the third feature data are combined in the order of timestamps to form the time series feature set.

4. An agricultural pest monitoring system as claimed in claim 2, characterized in that: The pest monitoring result data includes the types of pests, the number and distribution of each pest; The pest monitoring result data calculated by the classification calculation model also includes the degree of pest activity.

5. The agricultural pest monitoring system according to claim 3, characterized in that: The auxiliary acquisition module is at least one of the following sensors: a sound sensor, an infrared sensor, and a vibration sensor.

6. A method for monitoring agricultural pests, characterized in that: Using an agricultural pest monitoring system as described in any one of claims 1 to 5, a plurality of monitoring points are set in a target area, including: Step A1, real-time collection of pest activity data at each monitoring point; Step A2, processing the pest activity data to obtain pest monitoring result data of the monitoring point, and obtaining the pest monitoring result data of the target area according to the pest monitoring result data; Step A3, storing the pest activity data and the pest monitoring result data; The pest activity data includes pressure data and image data, and the step A2 includes: Step A201, receiving the pressure data of the monitoring point and forming first time series data according to the timestamp, receiving the image data and forming second time series data according to the timestamp; Step A202, processing the first time series data to extract first characteristic data, and processing the second time series data to extract second characteristic data; Step A203: for each monitoring point, merging the first feature data and the second feature data in the order of timestamps to form a time series feature set; Step A204, using the time series feature set as input, using a clustering algorithm model to perform clustering calculation to obtain the pest monitoring result data of the monitoring point; Step A205, according to the layout principle and coverage area of ​​the monitoring points in the target area, the pest monitoring result data of the monitoring points calculated by the clustering algorithm model are processed by a systematic sampling calculation method to obtain the pest monitoring result data of the target area; The pressure data includes acquisition timestamp, force point, force direction, and force magnitude; The first characteristic data includes the pest's moving speed, moving direction, each movement stop time, and the change value of the force direction; The second characteristic data includes the movement trajectory of the pests.

7. The agricultural pest monitoring method according to claim 6, characterized in that: The step A2 further comprises: Step A206, sending the time series feature set to a remote database for storage; Step A207, calling the time series feature set within a preset time period from the remote database; Step A208, taking the time series data set within the preset time period as input, using the trained classification calculation model to perform classification calculation, and obtaining the pest monitoring result data of the monitoring point within the preset time period; Step A209, based on the layout principle and coverage area of ​​the monitoring points in the target area, a systematic sampling calculation method is used to process the pest monitoring result data of the monitoring points calculated by the classification calculation model to obtain the pest monitoring result data of the target area.

Images