Pesticide application system based on pesticide resistance monitoring

Abstract

The invention provides a pesticide application system based on pesticide resistance monitoring, and relates to the field of agricultural intelligent prevention and control, and the system comprises a pesticide resistance monitoring module, a pest image recognition module, a pesticide application decision analysis module, an intelligent spraying control module, a data acquisition and communication module, and a cloud database and model module. The system generates a resistance coefficient according to a resistance detection result of target pests or pathogens, calculates a recommended pesticide application concentration by using a multi-factor fusion model in combination with crop image recognition, a growth period, environmental parameters and historical pesticide application information, and controls pesticide application equipment to execute variable spraying. The system has prevention effect feedback and model self-learning functions, and an intelligent closed-loop pesticide application process of'detection-identification-decision-execution-feedback 'is realized. The method can significantly improve the pesticide use efficiency, delay the resistance evolution and reduce the environmental risk, is suitable for precise prevention and control scenes of various crop diseases and insect pests, and has a good agricultural application prospect.

Description

Technical Field

[0001] This invention relates to the field of intelligent agricultural pest control technology, and more specifically, to a pesticide application system based on pesticide resistance monitoring. Background Technology

[0002] In modern agricultural production, pesticide application remains one of the core means of controlling crop diseases and pests and ensuring crop yield and quality. However, with the large-scale and continuous use of pesticides, the resistance of pests and pathogens has intensified year by year, and traditional application methods have gradually exposed serious limitations. On the one hand, farmers often rely on experience when applying pesticides, lacking real-time judgment of the resistance level of pests and diseases, which easily leads to inappropriate pesticide selection and unbalanced dosage settings, resulting in control failure. On the other hand, blindly increasing the dosage of pesticides can accelerate the evolution of resistance and trigger a chain of problems such as pesticide residues, environmental pollution, and ecological imbalance.

[0003] Pesticide resistance is an adaptive defense mechanism that pests and pathogens gradually develop under long-term pesticide selection pressure through mechanisms such as gene mutation, enhanced expression of metabolic enzymes, or changes in target conformation. Numerous studies have shown that once resistance develops, pesticides within the traditional dosage range become ineffective, and resistance levels rise rapidly with increasing application frequency. While some research institutions and government monitoring stations are currently conducting resistance testing, problems remain, including limited coverage, slow response times, and data fragmentation, making it difficult to provide effective guidance for farmers' daily pesticide application practices.

[0004] In recent years, with the development of agricultural informatization and intelligentization, some crop disease and pest diagnosis and variable-rate pesticide application systems have emerged, initially possessing regional identification and automatic control capabilities. However, existing systems generally lack dynamic perception of pesticide resistance and intelligent adjustment mechanisms for pesticide application strategies, failing to achieve a closed-loop application process of "perception-decision-control-feedback." Especially in complex scenarios involving multiple pesticides, coexistence of multiple diseases and pests, and crops at different growth stages, current systems still exhibit significant shortcomings in data fusion, model judgment, and precise execution.

[0005] Therefore, it is urgent to build a pesticide application system based on pesticide resistance monitoring in order to achieve the green governance goal of efficient, low-residue, and delayed resistance in pests and diseases. Summary of the Invention

[0006] The purpose of this invention is to solve the problems of the prior art mentioned in the background section and to provide a pesticide application system based on pesticide resistance monitoring.

[0007] The above-mentioned objective of the present invention is achieved as follows:

[0008] A pesticide application system based on pesticide resistance monitoring includes the following modules:

[0009] Pesticide resistance monitoring module: used to obtain resistance information of target pests or pathogens to a variety of commonly used pesticides. It calculates the corresponding resistance coefficient R by detecting resistance gene expression, lethal dose, and survival rate indicators, and uses this coefficient to make pesticide application strategy decisions.

[0010] Pest and disease image recognition module: used to collect field image data through crop image acquisition equipment, and use image recognition model to extract information such as pest and disease type, distribution location, and degree of damage, forming a pest and disease recognition matrix I(x,y,t);

[0011] The pesticide application decision analysis module is used to calculate the recommended pesticide concentration C by combining data such as resistance coefficient R, identification matrix I(x,y,t), historical pesticide application data H(t), environmental factors S, and crop growth stage G(t). opt The recommended concentration is calculated using the following formula:

[0012]

[0013] Among them, C opt The recommended application concentration is given by E, the target expected control efficacy, R, the resistance coefficient (range: (0,1), where a smaller value indicates stronger resistance), S, the environmental sensitivity factor (considering the effects of temperature, humidity, and wind speed, with a value range: (0,1), G(t) (crop growth stage coefficient, e.g., 1.0 for seedling stage, 0.8 for flowering stage, and 0.6 for grain-filling stage), H(t) (historical residual concentration), and θ (residual effect superposition coefficient, with a value range: 0-1).

[0014] Intelligent spray control module: includes variable spray control unit, spray path planning unit and operation execution unit, used to control the spraying according to C. opt The generated regional pesticide concentration distribution map controls the application of pesticides by the application equipment as needed;

[0015] Data acquisition and communication module: used to collect sensor information and upload field monitoring data, identification results and pesticide application records to the cloud system via 4G / 5G / LoRa / NB-IoT networks;

[0016] Cloud-based database and model module: Used to store historical pest and disease images, resistance lineages, operation trajectories, and pesticide application feedback data, supporting iterative model training and intelligent recommendation optimization.

[0017] As a preferred embodiment of the present invention, the pesticide resistance monitoring module includes a sample processing unit, a rapid resistance detection unit, and a data conversion unit. The rapid detection unit performs quantitative resistance analysis on target pests or pathogens based on quantitative real-time PCR, bioassay, or electrochemical sensing technology, and maps the results to a resistance coefficient R according to the following rules:

[0018] High resistance (survival rate 280%) is mapped to R = 0.2;

[0019] The intermediate antibody level (survival rate 40-80%) is mapped to R = 0.5;

[0020] Sensitivity (survival rate <40%) is mapped to R=1.0.

[0021] As a preferred technical solution of the present invention, the image recognition module includes an image acquisition device, a multimodal image preprocessing unit, and a target detection network based on deep learning. The recognition matrix I(x,y,t) is a two-dimensional spatiotemporal matrix, representing the pest density value at time t and coordinate (x,y), which is used for subsequent variable spraying control.

[0022] As a preferred technical solution of the present invention, the decision analysis module realizes intelligent pesticide concentration calculation based on the following steps:

[0023] (1) Extract the target block I(x,y,t) from the identification module;

[0024] (2) Read the current crop growth stage G(t);

[0025] (3) Query the resistance coefficient R of the pesticide corresponding to this crop;

[0026] (4) Obtain factors S from environmental sensors, such as temperature, humidity, and wind speed;

[0027] (5) The residual effect is corrected by combining the historical application concentration H(t).

[0028] (6) Execute the formula: Output the recommended application concentration.

[0029] As a preferred technical solution of the present invention, the intelligent spraying control module includes:

[0030] The block application concentration matching unit is used to match C opt Assigned to spatial points in the recognition matrix;

[0031] The trajectory optimization unit is used to plan the spraying path based on the RTK precision positioning system and crop arrangement.

[0032] The spray control actuator is used to adjust the nozzle flow rate in real time to achieve variable-rate pesticide application.

[0033] As a preferred technical solution of the present invention, the data communication module supports bidirectional communication with UAVs, ground operation machines, and mobile terminals, enabling real-time data uploading, remote command issuance, and operation status feedback.

[0034] As a preferred technical solution of the present invention, the cloud model module includes an image recognition training model, a drug concentration optimization model, and a resistance evolution trend prediction model, wherein the image recognition adopts a convolutional neural network, and the concentration prediction adopts a hybrid algorithm of random forest and Bayesian network.

[0035] As a preferred technical solution of the present invention, the system can construct a time series prediction model based on historical resistance data over multiple time periods, predict the trend of resistance changes, automatically generate a resistance risk curve, and use it to optimize subsequent detection frequency and application strategies.

[0036] As a preferred technical solution of the present invention, the system automatically collects the actual pesticide concentration C of the spraying equipment after pesticide application. real Feedback on field control efficacy F res The efficacy deviation index and concentration error index were calculated as follows:

[0037] Prevention effectiveness achievement rate:

[0038] Concentration deviation value: ΔC=|C red -C opt |;

[0039] Where: η is the efficacy achievement rate, 0≤η≤1, ΔC is the difference between the recommended concentration and the actual concentration, in mg / L.

[0040] As a preferred technical solution of the present invention, the system is applicable to multiple crop types such as rice, wheat, corn, cotton and vegetables. It can automatically match the corresponding pest and disease identification model, pesticide resistance database and crop growth period parameter table according to the crop code, so as to realize the sharing and automatic adaptation of a single system for multiple crops.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] This invention integrates rapid pesticide resistance monitoring technology, enabling real-time assessment of the resistance levels of pests and pathogens in the field, and quantifying these resistance levels into resistance coefficients for pesticide application decision-making. Compared to traditional experience-based pesticide application methods, this system can dynamically adjust pesticide types and concentrations based on actual resistance status, avoiding both ineffective application and excessive pesticide use. This significantly improves control efficiency while effectively delaying the formation and spread of pest resistance, demonstrating significant value for crop protection and ecological security.

[0043] This invention introduces a variable-rate pesticide application model based on image recognition and multi-factor fusion reasoning. This model combines multi-source data such as crop growth stage, environmental parameters, historical residues, and pest and disease density to calculate precise pesticide concentrations and guide intelligent equipment to implement differentiated spraying in different areas. This method can significantly reduce the total amount of pesticides used, avoid overspraying and missed spraying, improve pesticide utilization, reduce pollution to soil, water bodies, and non-target organisms, and promote green and sustainable agricultural development.

[0044] This invention establishes a closed-loop feedback mechanism for operational data, comparing actual spraying trajectories and efficacy results with model-recommended values. It evaluates application effectiveness in real time and corrects subsequent parameters, thereby enabling continuous self-learning and optimization updates of the model. The system supports remote cloud management and adaptability to multiple crops and scenarios, possessing excellent intelligence, modularity, and scalability. It is particularly suitable for precision agriculture, green pest control demonstration zones, and high-quality development scenarios in modern agriculture, demonstrating significant industrial application prospects and socio-economic benefits. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0046] Figure 1 This is a system block diagram of a pesticide application system based on pesticide resistance monitoring.

[0047] Figure 2 This is a diagram showing the composition of an intelligent spraying unit in a pesticide application system based on pesticide resistance monitoring. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0049] The following combination Figures 1-2 The specific embodiments of the present invention will be described in detail below.

[0050] This invention proposes a pesticide application system based on pesticide resistance monitoring. The system integrates multiple sub-modules, including resistance detection, image recognition, intelligent decision-making, precision application, and feedback learning, enabling scientific control and real-time adjustment of the pesticide application process. The system mainly includes a pesticide resistance monitoring module, a pest and disease image recognition module, a pesticide application decision analysis module, an intelligent spraying control module, a data acquisition and communication module, and a cloud database and model module. These modules collaborate and transmit commands via wired or wireless means, forming a complete closed-loop system for precise pesticide use.

[0051] During field operations, the pesticide resistance monitoring module first performs rapid detection of target pest and disease samples. This module can perform quantitative analysis of resistance genes or lethal doses through quantitative real-time PCR bioassay or nanosensoring. The system converts the detection results into a standardized resistance coefficient R, where sensitive pests and diseases correspond to R=1.0, moderately resistant pests and diseases correspond to R=0.5, and highly resistant pests and diseases correspond to R=0.2. This parameter is used as one of the important inputs to the decision-making model.

[0052] Meanwhile, the pest and disease image recognition module uses a drone or orbital platform equipped with a multispectral imaging device to collect images. The deep convolutional neural network deployed on the terminal or cloud automatically analyzes the image data to identify the types, distribution locations and densities of pests and diseases, and generates an identification matrix I(x,y,t), where x and y represent geographical coordinates, t represents time, and the matrix value reflects the intensity of pest and disease occurrence within a unit area.

[0053] The image recognition results were then used to generate pest and disease heat maps and serve as an important basis for variable-based pesticide application decisions. Based on the resistance coefficient and image recognition matrix, the pesticide application decision analysis module further retrieved the crop growth stage parameter G(t) of the current field, the environmental sensitivity factor S mapped from meteorological monitoring data (temperature, humidity, wind speed, etc.), and the pesticide residue concentration H(t) from historical application records. The recommended application concentration C was then calculated using a multi-factor weighted model. opt The calculation follows the formula:

[0054]

[0055] Wherein, E represents the expected control efficacy, usually set to 90 or 95; R is the resistance coefficient of the target crop, provided by the resistance monitoring module; S is the environmental sensitivity factor, ranging from 0 to 1, used to correct the utilization efficiency of pesticides under different meteorological conditions; G(t) is the crop growth stage coefficient, set according to the sensitivity of different crops to pesticides at different stages such as seedling stage, jointing stage, and heading stage; H(t) is the residual concentration after the last application or several consecutive applications; θ is the residual effect superposition correction coefficient, representing whether the current concentration should consider the influence of the previous residual effect, the specific value depends on the crop type and application frequency, and is usually set between 0 and 1.

[0056] Taking a typical cotton field pesticide application process as an example, detection showed that cotton bollworms had moderate resistance to a certain pyrethroid pesticide (R=0.5). The image recognition module identified areas with high-density insect populations exceeding 30%. The current field temperature was 31℃, humidity was 68%, and wind speed was 2.0m / s. Based on this, the system calculated an environmental sensitivity factor S=0.8. The cotton is currently in the budding stage and is relatively sensitive to pesticides, so G(t)=0.9 was set. The historical residual concentration of the previous application was H(t)=4mg / L, the residual effect superposition coefficient was set to θ=0.5, and the expected control efficacy was E=90. After substituting into the formula, the system automatically outputs the recommended application concentration:

[0057]

[0058] This concentration value will be transmitted to the intelligent spraying equipment via the spraying control module. The equipment will apply pesticides according to the pest and disease density in different areas of the identification matrix, i.e., the recommended concentration of 252 mg / L will be used in high-density areas, and the dosage will be reduced proportionally in low-density areas. The path planning system will automatically generate an efficient spraying trajectory by combining crop layout information and obstacle distribution, while controlling the nozzle flow rate and angle in real time to ensure accurate, efficient, and low-loss operation.

[0059] Throughout the operation, the data acquisition and communication module continuously collects operation parameters (such as spray concentration, equipment speed, spray trajectory, etc.) and environmental data (temperature, humidity, wind speed, etc.), and uploads all data synchronously to the cloud database.

[0060] The cloud platform records the entire operation process and updates the prediction model by combining actual field feedback data. The system compares the actual control efficacy F after pesticide application. res The ratio of the efficacy value to the originally set efficacy value E is used to calculate the efficacy achievement rate. And the deviation in pesticide application concentration: ΔC=|C real -C opt | is used to determine the execution error and model reliability, and accordingly corrects the recommended application concentration and application method for the next application, forming a continuously optimized intelligent feedback mechanism.

[0061] The system of the present invention is applicable to a variety of crops, including rice, corn, wheat, cotton, tomatoes, peppers, etc. By replacing the image recognition model and the resistance database, it can achieve automatic adaptation and cross-regional promotion of multiple pests, multiple pesticides, and multiple crops, and has good generality, scalability, and industrialization prospects.

[0062] The following is a detailed explanation with specific embodiments: A certain plot is a summer cotton planting field, covering an area of about 35 mu. Recently, cotton bollworms have continuously occurred, and there are traces of leaf biting and egg deposition spots of varying degrees in the field. To prevent the spread of pests and ensure the yield, the farmer starts the "pesticide application system based on pesticide resistance monitoring" described in the present invention for prevention and control operations.

[0063] At 8:00 am on the day of operation, several pest samples are collected by the field patrol personnel through the portable terminal and put into the portable resistance detection device supporting this system. This device uses the fluorescence quantitative PCR method to quickly detect the expression levels of resistance genes related to pesticide metabolism such as CYP450 and GSTs in the pest bodies. The detection results show that the target pests have obvious moderate resistance to the commonly used cypermethrin. The system automatically maps the resistance level to the resistance coefficient R = 0.5 and uploads it to the cloud model module.

[0064] At the same time, the drone equipped with a multispectral camera starts to conduct aerial photography on the target field. The flight altitude is set at 25 meters, and the resolution is 5 cm / pixel. The image recognition module analyzes the aerial photography images through the deployed convolutional neural network model (ResNet50-FPN), identifies multiple areas with dense pest populations, and the average pest density estimate is 6 per square meter. The system outputs the pest and disease identification matrix I(x, y, t), where the high-risk area covers about 28% of the total field area. Subsequently, the system automatically calls the current meteorological station data and combines the feedback from the on-site micro meteorological sensors: the current temperature is 31°C, the relative humidity is 68%, and the wind speed is 2.1 m / s. The system generates the environmental sensitivity factor S = 0.8 according to the empirical rule.

[0065] The crop is currently in the flowering stage, and the system calls the parameter setting G(t) = 0.85 from the crop database. It is found from the database that the last pesticide application was 6 days ago, and the residual concentration estimate is H(t) = 3.5 mg / L. The residual effect correction coefficient is set to θ = 0.5. The user sets the expected control efficacy E = 90 at the terminal.

[0066] The system automatically executes the pesticide application concentration calculation formula:

[0067]

[0068] The intelligent spraying module randomly reads the recognition matrix and, in conjunction with the recommended concentration, generates a variable spraying plan: the spraying concentration is set at 266.5 mg / L for areas with high insect density, 180 mg / L for medium-risk areas, and 100 mg / L for low-risk areas; the drone takes off automatically to execute the mission based on the planned path, speed, and spray width control.

[0069] The entire operation lasted approximately 37 minutes, covering an area of ​​approximately 34.6 acres. The data acquisition and communication module uploaded spraying logs in real time, including actual flow rate, concentration records, equipment flight path, and environmental changes. Five days after the operation, an inspection revealed that the insect population density in the target area had decreased to less than 0.5 insects per square meter, with a control efficacy assessment of 93%. The system automatically calculated the control efficacy achievement rate as follows:

[0070]

[0071] The actual spray concentration fluctuation range is recorded as C. real =270±5mg / L, deviation value:

[0072]

[0073] The actual spray concentration fluctuation range is recorded as C. real =270±5mg / L, deviation value:

[0074] ΔC=|C real -C opt |=|270-266.46|=3.54mg / L

[0075] The system considers the model prediction error to be within an acceptable range (≤5mg / L) and updates the residual effect superposition factor and environmental factor weights of the cloud model.

[0076] This embodiment demonstrates that the system described in this invention can realize a complete process of "rapid resistance detection + intelligent pest and disease identification + multi-factor reasoning decision-making + precise application of pesticides based on regional variables + closed-loop feedback of effects" in actual farmland scenarios, and has good practicality, intelligence and field adaptability.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A pesticide application system based on pesticide resistance monitoring, characterized in that, Includes the following modules: Pesticide resistance monitoring module: used to obtain resistance information of target pests or pathogens to a variety of commonly used pesticides. It calculates the corresponding resistance coefficient R by detecting resistance gene expression, lethal dose, and survival rate indicators, and uses this coefficient to make pesticide application strategy decisions. Pest and disease image recognition module: used to collect field image data through crop image acquisition equipment, and use image recognition model to extract information on pest and disease type, distribution location and degree of damage, forming a pest and disease recognition matrix I(x,y,t); The pesticide application decision analysis module is used to calculate the recommended pesticide concentration C by combining the resistance coefficient R, the identification matrix I(x,y,t), historical pesticide application data H(t), environmental factors S, and crop growth period data G(t). opt The recommended concentration is calculated using the following formula: Among them, C opt The recommended application concentration is given by E, the target expected control efficacy, R, the resistance coefficient (range: (0,1), where a smaller value indicates stronger resistance), S, the environmental sensitivity factor (considering the effects of temperature, humidity, and wind speed, with a value range: (0,1), G(t) (crop growth stage coefficient, e.g., 1.0 for seedling stage, 0.8 for flowering stage, and 0.6 for grain-filling stage), H(t) (historical residual concentration), and θ (residual effect superposition coefficient, with a value range: 0-1). Intelligent spray control module: includes variable spray control unit, spray path planning unit and operation execution unit, used to control the spraying according to C. opt The generated regional pesticide concentration distribution map controls the application of pesticides by the application equipment as needed; Data acquisition and communication module: used to collect sensor information and upload field monitoring data, identification results and pesticide application records to the cloud system via 4G / 5G / LoRa / NB-IoT networks; Cloud-based database and model module: Used to store historical pest and disease images, resistance lineages, operation trajectories, and pesticide application feedback data, supporting iterative model training and intelligent recommendation optimization.

2. The pesticide application system based on pesticide resistance monitoring according to claim 1, characterized in that, The pesticide resistance monitoring module includes a sample processing unit, a rapid resistance detection unit, and a data conversion unit. The rapid detection unit performs quantitative resistance analysis on target pests or pathogens based on quantitative real-time PCR, bioassays, or electrochemical sensing technology, and maps the results to a resistance coefficient R according to the following rules: High resistance (survival rate 280%) is mapped to R = 0.2; The intermediate antibody level (survival rate 40-80%) is mapped to R = 0.5; Sensitivity (survival rate <40%) is mapped to R=1.

0.

3. The pesticide application system based on pesticide resistance monitoring according to claim 1, characterized in that, The image recognition module includes an image acquisition device, a multimodal image preprocessing unit, and a target detection network based on deep learning. The recognition matrix I(x,y,t) is a two-dimensional spatiotemporal matrix, representing the pest density value at time t and coordinate (x,y), which is used for subsequent variable spraying control.

4. The pesticide application system based on pesticide resistance monitoring according to claim 1, characterized in that, The decision analysis module calculates intelligent pesticide application concentration based on the following steps: (1) Extract the target block I(x,y,t) from the identification module; (2) Read the current crop growth stage G(t); (3) Query the resistance coefficient R of the pesticide corresponding to this crop; (4) Obtain factors S from environmental sensors, such as temperature, humidity, and wind speed; (5) The residual effect is corrected by combining the historical application concentration H(t); (6) Execute the formula: Output the recommended application concentration.

5. The pesticide application system based on pesticide resistance monitoring according to claim 1, characterized in that, The intelligent spraying control module includes: The block application concentration matching unit is used to match C opt Assigned to spatial points in the recognition matrix; The trajectory optimization unit is used to plan the spraying path based on the RTK precision positioning system and crop arrangement. The spray control actuator is used to adjust the nozzle flow rate in real time to achieve variable-rate pesticide application.

6. The pesticide application system based on pesticide resistance monitoring according to claim 1, characterized in that, The data communication module supports two-way communication with drones, ground operation machines, and mobile terminals, enabling real-time data uploading, remote command issuance, and operation status feedback.

7. The pesticide application system based on pesticide resistance monitoring according to claim 1, characterized in that, The cloud-based model module includes an image recognition training model, a drug concentration optimization model, and a resistance evolution trend prediction model. The image recognition uses a convolutional neural network, and the concentration prediction uses a hybrid algorithm of random forest and Bayesian network.

8. The pesticide application system based on pesticide resistance monitoring according to claim 1, characterized in that, The system can build a time series prediction model based on historical resistance data over multiple time periods, predict resistance change trends, automatically generate resistance risk curves, and be used to optimize subsequent detection frequency and application strategies.

9. The pesticide application system based on pesticide resistance monitoring according to claim 1, characterized in that, The system automatically collects the actual pesticide concentration C of the spraying equipment after application. real Feedback on field control efficacy F res The efficacy deviation index and concentration error index were calculated as follows: Prevention effectiveness achievement rate: Concentration deviation value: ΔC=|C real -C opt |; Where: η is the efficacy achievement rate, 0≤η≤1, ΔC is the difference between the recommended concentration and the actual concentration, in mg / L.

10. The pesticide application system based on pesticide resistance monitoring according to claim 1, characterized in that, The system is applicable to various crop types such as rice, wheat, corn, cotton, and vegetables. It can automatically match the corresponding pest and disease identification model, pesticide resistance database, and crop growth period parameter table according to the crop code, realizing the sharing and automatic adaptation of a single system for multiple crops.

Images