A tobacco field pest and disease monitoring method and system based on image recognition

Through the tobacco field pest monitoring method based on image recognition, the problem of difficult to quantitatively identify pest resistance in the prior art is solved by calculating the number and range of pests and diseases, and the accurate monitoring and timely feedback of drug resistance is achieved, and pests and diseases are improved. Prevention and treatment effect.

CN118781550BActive Publication Date: 2024-11-19LIANGSHAN YI AUTONOMOUS PREFECTURE ACAD OF AGRI SCI +1
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202411217468.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-02
Publication Date
2024-11-19
Estimated Expiration
2044-09-02

AI Technical Summary

Technical Problem

The prior art is difficult to directly calculate and identify drug resistance of pests and diseases, resulting in inaccurate monitoring of pests and diseases and difficult to deal with drug resistance problems in a timely manner.

Method used

The tobacco field pest monitoring method is used based on image recognition, and the drug resistance of pests and diseases is indirectly quantified through relevant calculations of the active number and active range of pests and diseases, including classification of pests and diseases, determining the area to be monitored, calculating activity and drug resistance.

Benefits of technology

It improves the accuracy and timeliness of drug resistance in pests and diseases, reduces the difficulty of pest sampling, reduces the amount of image recognition calculation, enhances the evaluation factors for drug resistance, and can more accurately reflect the resistance of pests and diseases to chemical agents.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118781550B_ABST
    Figure CN118781550B_ABST
Patent Text Reader

Abstract

The present invention relates to the technical field of pest and disease resistance monitoring, and in particular to a tobacco field pest and disease monitoring method and system based on image recognition. The method comprises: sampling a tobacco field, obtaining an initial picture of a sample plot, inputting the initial picture into a pest and disease classification model, classifying the pest and disease according to the appearance characteristics of the pest and disease, and dividing the pest and disease into monitoring categories according to the type and appearance characteristics of the pest and disease; re-determining the area to be monitored of the pest and disease according to the appearance characteristics and the monitoring category, re-shooting the area to be monitored, and obtaining the picture to be monitored; obtaining the movement position of the pest and disease in the picture to be monitored, and determining the active area of ​​the pest and disease based on multiple movement positions; calculating the activity of the pest and disease according to the geometric size of the active area; calculating the pesticide resistance of the first sample plot according to the activity recorded in the historical record and the current activity; using the pesticide resistance of the first sample plot to estimate the pesticide resistance of the tobacco field, and realizing the timely monitoring of the pesticide resistance of the tobacco field pest and disease in the actual scene.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of pest and disease resistance monitoring, and in particular to a tobacco field pest and disease monitoring method and system based on image recognition. Background Art

[0002] At present, the prevention and control measures for pests and diseases are still mainly chemical control. However, in the actual prevention and control process, the effect of pest and disease control is limited by the farmers' insufficient professional knowledge reserves in the agricultural field and the impact of long-term use of chemical control. If the same type of chemical control measures are used for a long time, pests and diseases will often develop resistance to chemical control, making it difficult to maintain the inhibitory effect of chemical control on pests and diseases for a long time. When the inhibitory effect of pests and diseases decreases, farmers often directly increase the concentration of chemical control, resulting in further enhancement of the resistance of pests and diseases, which makes the control effect of pesticides decrease instead of increase, and even causes the chemical residues in crops to exceed the standard.

[0003] With the rapid development of global agricultural production, the problem of crop diseases and pests is becoming increasingly serious. The impact of pest resistance has posed a major threat to agricultural production safety and crop yields. Traditional pest and disease monitoring methods mainly rely on manual inspections and empirical judgments. This method is not only inefficient, but also has a high misjudgment rate and is difficult to detect pests and diseases in real time. In recent years, with the rapid development of technologies such as the Internet of Things, big data, image recognition, and artificial intelligence, their application in the agricultural field has also made significant progress. Pest and disease monitoring technology has gradually developed in the direction of intelligence and automation, providing strong support for pest and disease prevention and control in agricultural production.

[0004] Since drug resistance is a qualitative indicator, it is difficult for technologies such as image recognition to directly quantify and identify drug resistance at the gene level. Currently, it is a feasible way to measure the number of deaths under the influence of chemical agents. For example, the application document with application number CN202010906528.0 provides a method for quickly identifying the resistance of fall armyworm to diamide insecticides. The fall armyworm obtained from the field is cultured in the laboratory, and then the fall armyworms that died under five different concentrations of chemical agents are counted. The resistance is obtained by drawing a toxicity regression curve based on the lethal chemical concentration. However, since it is difficult to achieve consistency in the degree of resistance of surviving pests and diseases, it is difficult to accurately reflect the resistance based on the number of survivors alone. Therefore, the present application provides a tobacco field pest and disease monitoring method and system based on image recognition. Through the relevant calculation results of the active areas of tobacco field pests and diseases, the resistance of pests and diseases is indirectly quantified, which improves the accuracy of monitoring resistance and realizes the timely monitoring of tobacco field pest and disease resistance in actual scenarios. Summary of the invention

[0005] The purpose of the present invention is to provide a tobacco field pest and disease monitoring method and system based on image recognition in view of the problem that current image recognition and other technologies are difficult to directly perform quantitative calculation and identification of pesticide resistance. By calculating the relevant results of the active number and active range of tobacco field pests and diseases, the pesticide resistance of pests and diseases is indirectly quantified, thereby realizing timely monitoring of the pesticide resistance of tobacco field pests and diseases.

[0006] In order to achieve the above-mentioned object of the invention, the present invention provides the following technical solutions:

[0007] A tobacco field pest and disease monitoring method based on image recognition, comprising the following steps:

[0008] Sampling the tobacco field to obtain a first sample plot; obtaining an initial image of the first sample plot, inputting the initial image into a pest classification model, classifying the pests according to their appearance characteristics, and dividing the pests into monitoring categories according to their types and appearance characteristics;

[0009] Re-determine the area to be monitored for the pests and diseases according to the appearance characteristics and the monitoring category, and re-photograph the area to be monitored to obtain a picture to be monitored;

[0010] When the preset time step is reached, different movement positions of the pests in the image to be monitored are obtained, and the active area of ​​the pests is determined based on the multiple movement positions; the activity of the pests is calculated according to the geometric size of the active area;

[0011] The pesticide resistance of the first plot is calculated according to the activity degree recorded in the history and the current activity degree; and the pesticide resistance of the tobacco field is estimated using the pesticide resistance of the first plot.

[0012] As a preferred technical solution of the present application, the method of classifying the pests and diseases into monitoring categories according to the types of pests and diseases and the appearance characteristics thereof includes:

[0013] Determine the size of the pests and diseases according to the appearance characteristics; determine the movement mode of the pests and diseases according to the type of the pests and diseases and the appearance characteristics; preset a body size threshold; compare the body size of the pests and diseases with the body size threshold, and classify the pests and diseases into monitoring categories according to the comparison result and the movement mode.

[0014] As a preferred technical solution of the present application, the method of determining the movement mode of the pests and diseases according to the type of the pests and diseases and the appearance characteristics includes:

[0015] The appearance features include contour features and movement organ features, and the movement organ features include flight organ features and foot organ features; the movement mode of the pests is determined based on the type of the pests and the movement organ features.

[0016] As a preferred technical solution of the present application, the step of calculating the activity of the pests and diseases according to the geometric size of the active area includes:

[0017] Obtain active areas of multiple different clusters of the pests and diseases when the pests and diseases are not affected by any chemical agents, calculate the geometric sizes of the active areas, and determine a benchmark value based on the multiple geometric sizes; take the ratio of the geometric size of the active area at that time to the benchmark value as the activity level of the pests and diseases.

[0018] As a preferred technical solution of the present application, the calculation of the drug resistance of the first sample according to the activity recorded in the history and the current activity includes:

[0019] Obtain the activity of the monitored area before using any chemical agent, recorded as the initial activity; obtain the activity of the monitored area after using chemical agents for the first time, recorded as the first activity; obtain the activity of the monitored area after using chemical agents for the second time, recorded as the second activity; calculate the first difference between the second activity and the first activity, and take the ratio of the first difference to the initial activity as the drug resistance of the monitored area.

[0020] As a preferred technical solution of the present application, a size-to-area ratio threshold is preset, a first ratio of the size of the pests and diseases to the horizontal area of ​​the monitored area is calculated, the first ratio is compared with the size-to-area ratio threshold, and a method for calculating the active area is determined based on the comparison result and the monitoring category.

[0021] As a preferred technical solution of the present application, the calculation method of the active area includes: when reaching multiple time steps, obtaining the adjacent movement positions of a single or multiple pests and diseases, calculating the coordinate distance according to the coordinates of the movement positions, and summing the multiple coordinate distances to obtain the cumulative movement distance.

[0022] As a preferred technical solution of the present application, the method for calculating the active area includes: when reaching multiple time steps, obtaining multiple pictures to be monitored of a single or multiple pests and diseases, fusing the contour features of the pests and diseases in the multiple pictures to be monitored into one picture through an image fusion algorithm, extracting the contour features of the pests and diseases in the fused picture to obtain fused contour features; calculating the area of ​​the figure enclosed by the fused contour features to obtain the motion area.

[0023] As a preferred technical solution of this application, the calculation method of the active area also includes:

[0024] When a plurality of the time steps are reached, a plurality of top-view images and a plurality of side-view images of the area to be monitored are acquired;

[0025] Fusing the plurality of overhead images, drawing polygons for the first point set formed by the pests and diseases in the fused images, and obtaining a horizontal active area;

[0026] Fusing the multiple side-view images to obtain a second point set formed by the pests and diseases in the fused image, calculating the vertical distance between the pests and diseases at the highest position and the pests and diseases at the lowest position in the second point set, and obtaining the maximum vertical movement height;

[0027] The size of the motion space is obtained by calculating the product of the area of ​​the horizontal active region and the maximum vertical motion height.

[0028] The present application also provides a tobacco field pest and disease monitoring system based on image recognition, which is used to implement the pest and disease detection method, and includes the following modules:

[0029] A classification module is used to divide the tobacco field into sample plots; after selecting the first sample plot, an initial image of the first sample plot is received, the initial image is output to a pest classification model, the appearance features of the pests are extracted for classification, and a monitoring category is output for the pests according to the types of the pests and the appearance features;

[0030] A monitoring module, for receiving the appearance features and the monitoring category, adjusting the area to be monitored for the pests and diseases; re-photographing the area to be monitored, and outputting a picture to be monitored;

[0031] The activity module is used to receive the movement position of the pest in the image to be monitored when the preset time step is reached, determine the active area of ​​the pest based on multiple movement positions; calculate and output the activity of the pest according to the geometric size of the active area;

[0032] The pesticide resistance module is used to receive the activity recorded in the history and the current activity and output the pesticide resistance of the first sample plot; and estimate and output the pesticide resistance of the tobacco field according to the pesticide resistance of the first sample plot.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] 1. It is difficult to sample all pests and diseases in tobacco fields, and it takes a long time to sample. By sampling pests and diseases with a high concentration, the pesticide resistance of tobacco fields can be estimated based on the pesticide resistance of the sample plots, which is conducive to reducing the difficulty of sampling pests and diseases, reducing the amount of calculation in the image recognition process, achieving comprehensive and accurate statistics of pests and diseases in the sample plots, and improving the accuracy of calculating pesticide resistance.

[0035] 2. The movement position of the pests and diseases with a smaller active area in the first sample plot is difficult to be accurately identified. To address this problem, the present application presets different monitored areas for different pests and diseases, thereby increasing the proportion of the body size of small pests and diseases in the images to be detected, making the body size and movement trajectory of small pests and diseases easier to identify, which is beneficial to improving the accuracy of identifying the movement position of the pests and diseases through images, making it easier to track the pests and diseases.

[0036] 3. Since the size of larger pests and diseases has an impact on the size of the active area when calculating the active area, pests and diseases of different sizes are distinguished by setting a size threshold, and are graded in combination with the movement mode of the pests and diseases. Different image processing methods and different active area calculation methods are used for pests and diseases of different grades, which can reflect the impact of body size and movement mode on the active area and improve the accuracy of calculating the active area.

[0037] 4. The pesticide resistance of pests and diseases reflects the resistance of pests and diseases to chemical agents. The stronger the resistance, the less affected the pests and diseases are by the chemicals, and the higher the activity. When using chemicals in tobacco fields, the concentration of chemicals in different areas is different due to the influence of environmental factors such as temperature and wind force; at the same time, different pests and diseases have different resistance to chemicals, so the active areas of pests and diseases will also be different. Therefore, reflecting the activity of pests and diseases under different movement abilities and movement methods from multiple angles is conducive to the monitoring of pest and disease resistance. This application reflects the activity of pests and diseases from three levels: linear, planar, and three-dimensional by calculating the movement distance, movement area, and movement space of pests and diseases, and adds an evaluation factor for resistance, which is conducive to distinguishing different degrees of resistance.

[0038] 5. In order to solve the problem that the resistance of pests and diseases is difficult to quantify directly after using chemical agents, the movement position of pests and diseases in the first sample is obtained by image recognition, and then the activity of pests and diseases is calculated by the movement position. The resistance is quantified by the ratio of the change in activity caused by resistance to the activity of pests and diseases without using chemical agents. This is conducive to observing the effect of chemical agents after using chemical agents in the first sample, so as to adjust the subsequent pest and disease control measures in a targeted manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 Schematic diagram of the process for calculating insecticide resistance of pests and diseases;

[0040] Figure 2 This is a structural diagram of the pest and disease monitoring system. DETAILED DESCRIPTION

[0041] To make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be described clearly and completely in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them.

[0042] Therefore, the following detailed description of the embodiments of the present invention is not intended to limit the scope of the invention claimed for protection, but merely represents some embodiments of the present invention. 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.

[0043] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features and technical solutions in the embodiments may be combined with each other.

[0044] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0045] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "upper", "lower", etc. is based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the invention product is usually placed when in use, or the orientation or positional relationship commonly understood by those skilled in the art. Such terms are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0046] Example 1: See Figure 1 As shown, the present embodiment provides a tobacco field pest and disease monitoring method based on image recognition, comprising the following steps:

[0047] Sampling the tobacco field to obtain a first sample plot; obtaining an initial image of the first sample plot, inputting the initial image into a pest classification model, classifying the pests according to their appearance characteristics, and dividing the pests into monitoring categories according to their types and the appearance characteristics.

[0048] First, the tobacco field is sampled. It is difficult to count, identify or track all the pests and diseases in the entire tobacco field. For example, the leaves on a single plant are prone to overlap in a bird's-eye view, the leaves of mature plants are lush and block each other, and some types of pests and diseases hide on the back of the leaves, making it difficult to take pictures of some pests and diseases. Therefore, in this embodiment, the plant protection expert divides the sample plot according to the tobacco field. The size and position of the sample plot are determined according to the concentration of the pests and diseases expressed in quantity. The more the number, the higher the probability of the area of ​​the active area of ​​the pests and diseases increasing. A larger sample plot is selected accordingly. After the plant protection expert selects the first sample plot, the pesticide resistance of the first sample plot is calculated to achieve quantification and estimation of the pesticide resistance of the entire tobacco field. In this way, it is conducive to reducing the difficulty of sampling pests and diseases, reducing the amount of calculation in the image recognition process, and achieving comprehensive and accurate statistics of the number and type of pests and diseases in the sample plot, thereby improving the accuracy of the calculation results. The criteria for screening out the first sample plot include but are not limited to expert experience or the severity of pests and diseases in the previous year; the screening methods include but are not limited to the five-point sampling method and the fixed-distance sampling method. In this embodiment, it is preferred to select based on the standard deviation of the historical data of each sample plot, and the sample plot with a small standard deviation, that is, the pest and disease data fluctuates smoothly, is preferred.

[0049] The pictures of the first sample plot are obtained from different angles, including but not limited to the top view, side view and top view of the first sample plot, which are recorded as initial pictures. The initial pictures include pictures taken manually, pre-installed high-definition shooting equipment and pictures collected by drone aerial photography. The pest and disease classification model is trained by training the training set and test set that are manually classified and labeled. The classification process is to use a neural network to extract the appearance features of the pest and disease, classify and predict the appearance features through the trained neuron parameters in the output layer, and output the name of the species of the pest and disease. The use of a pre-trained pest and disease classification model is conducive to improving the efficiency of classification and reducing the workload of plant protection personnel.

[0050] Due to the different mechanisms of action of chemical agents, the types of pests and diseases that can be prevented and treated are limited. Therefore, in this embodiment, the target object for monitoring drug resistance is the pests and diseases in the instructions corresponding to the chemical agents. Therefore, it is necessary to obtain the appearance pictures and types of the corresponding pests and diseases to be prevented and treated from the chemical agent-pest and disease database according to the mechanism of action of the chemical agent, and record the type set of pests and diseases that can be prevented and treated by the chemical agent as the first pest and disease set. After applying chemical agents to the tobacco field, the initial picture is input into the pest and disease classification model for image recognition to extract the appearance features of the pests and diseases. According to the appearance features, the type set of all pests and diseases in the initial picture is identified and recorded as the second pest and disease set. The intersection of the first pest and disease set and the second pest and disease set is the target pest and disease set that needs to be monitored for drug resistance after using the chemical agent. In this way, the pest and disease monitoring method is targeted at the pests and diseases to be monitored, which helps to further reduce the amount of data processed by the pest and disease monitoring system and improve processing efficiency.

[0051] Determine the movement mode of the pests and diseases according to the type of the pests and diseases and the appearance characteristics; preset a body size threshold; compare the body size of the pests and diseases with the body size threshold, and classify the pests and diseases into monitoring categories according to the comparison result and the movement mode.

[0052] It takes a certain amount of time for chemical control to affect pests and diseases, and the length of time is affected by a variety of factors, including but not limited to the type of chemical agent, the type of pest and disease, and its sensitivity and resistance to the chemical agent. During this time, the motility of pests and diseases gradually decreases under the influence of the chemical agent, and when the duration of the drug effect just ends, the motility of pests and diseases reaches the lowest level.

[0053] Different types of pests and diseases have different movement abilities and modes of movement. The active areas of pests and diseases are affected by their own influencing factors, including but not limited to body size and mode of movement. The modes of movement of the pests and diseases include but are not limited to flying, jumping, crawling and creeping. Since the active areas of flying and jumping are usually larger than the active areas of crawling and creeping, and based on the size of the first square in this embodiment, the active areas of flying and jumping can be regarded as the same, and the active areas of crawling and creeping can be regarded as the same. In this embodiment, the modes of movement of the pests and diseases are uniformly divided into two categories, including flying and crawling, which helps to simplify the classification system, reduce the influencing factors when calculating the active areas, and help reduce the amount of calculation of image processing. Since the microorganisms that cause diseases are too small, it is difficult to use low-cost image recognition equipment to count the pathogens in the tobacco field environment. Therefore, the image recognition used in this embodiment is used to identify insects in the tobacco field. The activity of the pests and diseases is judged according to the aggregation degree and active area reflected by the number of pests and diseases, and then the resistance is calculated according to the change in the activity of the pests and diseases after chemical control.

[0054] Different movement modes result in different sizes of active areas of pests and diseases, and the movement mode is determined by the movement organs identified in the appearance features. The appearance features include contour features and movement organ features, and the movement organ features include flight organ features and foot organ features. The movement organ features of some pests and diseases are difficult to clearly identify. For example, the flight organ of the beetle is only extended when flying, and the flight organ is completely folded in the elytra when it stops moving; the snail's movement mode is creeping, and there is no movement organ feature for identification. Therefore, it is necessary to determine the movement mode of the pest and disease in combination with the type of pest and disease and the movement organ features in the appearance picture under different movement states. For example, combining the beetle and the wings in the picture when in motion to determine that the beetle's movement mode is flying; combining the snail and the absence of movement organ features in the appearance picture to determine that the snail's movement mode is crawling.

[0055] The appearance pictures need to include the appearance pictures of the pests and diseases when they are stationary and when they are moving, and the appearance pictures are stored in the chemical agent-pests and diseases database. The chemical agent-pests and diseases database is used to store the relevant information of the chemical agents and the pests and diseases, including but not limited to the type name of the chemical agent, the corresponding pests and diseases, and the appearance pictures corresponding to the pests and diseases. By identifying the appearance pictures of pests and diseases in different movement states, it is helpful to reduce the situation where the characteristics of moving organs in a stationary state are difficult to identify, and improve the classification accuracy of the movement mode of pests and diseases.

[0056] The size of the pest is determined based on the contour features. The resolution of the initial image is obtained, and the area of ​​each pixel is calculated in combination with the width and height of the image. The contour features of the pest are extracted from the initial image through an edge detection algorithm, and then the number of pixels of the image surrounded by the contour features is obtained. The product of the area of ​​the pixel and the number of pixels is the area occupied by the pest in the initial image, that is, the size of the pest.

[0057] Due to territorial awareness and / or resource competition, the larger the size of the pests and diseases, the lower the degree of aggregation, the larger the active area and the more affected by the pests and diseases' own movement ability. The smaller the pests and diseases, the higher the degree of aggregation. Since the active area is very small and easily affected by other similar pests and diseases during image recognition, the active area of ​​small pests and diseases in this embodiment is represented by the active area of ​​the small pest and disease group. Select multiple different types of pest and disease groups in the tobacco field, randomly collect multiple pests and diseases from the pest and disease groups, obtain statistical values ​​of the body sizes of these pests and diseases, and preset body size thresholds for the pests and diseases according to the statistical values; for example, calculate the average body size of these pests and diseases, and use the average as the body size threshold.

[0058] The pests and diseases are divided into monitoring categories according to the size and movement mode of the pests and diseases. For example, in this embodiment, the monitoring categories of the pests and diseases are divided into four categories: if the size of the pests and diseases is greater than the size threshold, and the movement mode is flying, the monitoring category of the pests and diseases is the first pests and diseases. If the size of the pests and diseases is greater than the size threshold, and the movement mode is crawling, the monitoring category of the pests and diseases is the second pests and diseases. If the size of the pests and diseases is less than or equal to the size threshold, and the movement mode is flying, the monitoring category of the pests and diseases is the third pests and diseases. If the size of the pests and diseases is less than or equal to the size threshold, and the movement mode is crawling, the monitoring category of the pests and diseases is the fourth pests and diseases.

[0059] Since the size of larger pests and diseases has an impact on the size of the active area when calculating the active area, pests and diseases of different sizes are distinguished by setting a size threshold, and are graded in combination with the movement mode of the pests and diseases. Different image processing methods and different active area calculation methods are used for pests and diseases of different grades, which can reflect the impact of body size and movement mode on the active area and improve the accuracy of calculating the active area.

[0060] The area to be monitored of the pests and diseases is re-determined based on the appearance characteristics and the monitoring category. The pests and diseases with different mobile organ characteristics have different active areas, so the picture sizes used for image recognition of different pests and diseases are also different. For example, the movement mode of Noctuidae pests and diseases is flying, and the active area in the first sample plot is larger than the active area of ​​the snail in the same period of time. They can move in the entire first sample plot, and the movement position and / or movement trajectory are easy to identify in the initial picture. However, the movement position of the snail in the initial picture is relatively concentrated, and the movement trajectory is difficult to accurately and effectively identify in the initial picture. Therefore, the area to be monitored for image recognition and calculation of the active area size of Noctuidae pests and diseases needs to be larger than the area to be monitored for snails. Snails require an area to be monitored with more local details and clearer details.

[0061] At the same time, the active areas of the pests and diseases of different sizes in the first sample plot are also different. For the second pest and disease, the third pest and disease and the fourth pest and disease, their active areas are difficult to cover the entire first sample plot. Due to the influence of the focal length, viewing angle and movement ability of the pest and disease of the shooting equipment, the image content presented by the pest and disease in the initial picture is relatively concentrated, and it is difficult to accurately and effectively track the trajectory of the pest and disease, and different pest and disease individuals are also easy to affect each other. Therefore, it is often difficult to accurately and clearly identify the movement positions of these pests and diseases in the initial picture, and it is necessary to carry out targeted image preprocessing of the identification pictures of these pests and diseases and / or re-shoot the identification pictures for these pests and diseases. In this embodiment, the area to be monitored of the pest and disease is set according to the monitoring category. After determining the monitoring category of the pests and diseases, a rectangular coordinate system is established based on the initial picture to determine the coordinates of the pests and diseases whose resistance needs to be monitored. For the first pest and disease, it is preferred to use the coordinates of the individual body of the pest and disease, and for the second pest and disease, the third pest and disease, and the fourth pest and disease, it is preferred to use the centroid coordinates of the geometric figure surrounded by the edge of the pest and disease group. The position of the area to be monitored is determined according to the position of the pest and disease in the initial picture, and then the picture to be monitored for identifying and tracking the pest and disease is re-photographed.

[0062] After obtaining the target pest and disease set, the preset screening criteria for the area to be monitored are obtained according to the monitoring category, and the screening criteria are determined by the person in charge of the first plot according to the average value of the historical data of the activity area of ​​the pest and disease. In this embodiment, the screening criteria include: the area to be monitored for the first pest and disease is the first plot; the area to be monitored for the second pest and disease is a single plant in the first plot; the area to be monitored for the third pest and disease is a single plant area, the single plant area includes the single plant and the space above the plant, and the volume of the single plant area is the product of the projection area of ​​the single plant and n times the height of the single plant; the area to be monitored for the fourth pest and disease is the leaves on the single plant.

[0063] The movement position of the pests and diseases with a smaller active area in the first sample plot is difficult to be accurately identified. To address this problem, the present application presets different monitored areas for different pests and diseases, thereby increasing the proportion of the body size of small pests and diseases in the images to be detected, making the body size and movement trajectory of small pests and diseases easier to identify, which is beneficial to improving the accuracy of identifying the movement position of the pests and diseases through images, making it easier to track the pests and diseases.

[0064] At the same time, it is also necessary to preset a time step, obtain the movement positions of all pests and diseases in the target pest and disease set from the monitored picture, determine the active area where the pests and diseases are active based on multiple movement positions; and calculate the activity of the pests and diseases according to the geometric size of the active area.

[0065] Chemical agents may cause the affected pests to die or become less active, resulting in the activity area of ​​the pests that have not developed resistance to the drugs being smaller than the activity area of ​​the pests that have developed resistance to the drugs. In this embodiment, the movement position of the pests in the area to be monitored is obtained by image recognition, and the active area of ​​the pests in the area to be monitored is obtained through monitoring for a period of time, and the activity of the pests is calculated based on the size of the active area. In this embodiment, an image tracking algorithm is used to construct a pest monitoring model, and the pest monitoring model is used to monitor the target pest set in the area to be monitored to obtain the movement position and movement trajectory of the pests.

[0066] The effective time of the chemical agent is obtained. When the effective time is reached, a picture of the area to be monitored of the pests and diseases within the preset monitoring time is obtained and recorded as the picture to be monitored. A rectangular coordinate system is established based on the area to be monitored, and the picture to be monitored is input into the pest monitoring model to extract the movement position of the pests and diseases. Within the preset monitoring time, a time step is preset. The time step is determined according to the average movement speed of the pests and diseases within the preset monitoring time after the last use of the chemical agent. The larger the average movement speed, the smaller the time step. The movement position is the coordinate set of the pests and diseases when the time step is reached. According to the coordinates of the multiple movement positions, the active area of ​​the pests and diseases is determined. The active area includes the cumulative movement distance, movement area and movement space of the pests and diseases. The movement distance is the cumulative movement distance of the pests and diseases in the preset monitoring time, the movement area is the area of ​​the active area of ​​the pests and diseases within the monitoring time, and the movement space is the spatial volume of the active area of ​​the pests and diseases within the monitoring time.

[0067] Preferably, the activity of the pests is calculated based on the geometric size of the active area, and the method includes: obtaining the active areas of the pests of multiple different clusters when the pests are not affected by any chemical agent, calculating the geometric size of the active area, determining a reference value based on the multiple geometric sizes, the reference value including but not limited to the mean, median and mode; and taking the ratio of the geometric size of the active area at that time to the reference value as the activity of the pests. Taking the ratio of the active area affected by the chemical agent to the active area not affected as the activity can intuitively reflect the degree of change in the activity ability of the pests after being affected by the drug.

[0068] The pesticide resistance of pests and diseases reflects the resistance of pests and diseases to chemical agents. The stronger the resistance, the less affected the pests and diseases are by the chemicals, and the higher the activity. When using chemicals in tobacco fields, the concentration of chemicals in different areas is different due to the influence of the environment such as temperature and wind force; at the same time, different pests and diseases have different resistance to chemicals, so the active areas of pests and diseases will also be different. Therefore, reflecting the activity of pests and diseases under different movement abilities and movement methods from multiple angles is conducive to the monitoring of pest and disease resistance. The present application reflects the activity of pests and diseases from three levels: linear, planar and three-dimensional space by calculating the movement distance, movement area and movement space of pests and diseases, and adds an evaluation factor for resistance, which is conducive to distinguishing different degrees of resistance.

[0069] After obtaining the activity of the pests and diseases, the pesticide resistance of the first plot is calculated according to the activity recorded in the history and the current activity; and the pesticide resistance of the tobacco field is estimated using the pesticide resistance of the first plot.

[0070] The pesticide resistance of pests and diseases refers to the phenomenon that, after a period of development, pests and diseases develop resistance to chemical agents that were originally able to effectively control the population, and the chemical agents have a selective effect on the pests and diseases. Since each time a chemical agent with the same mechanism of action is used, the impact on pests and diseases that have already developed resistance is reduced, and a selective effect is once again exerted on pests and diseases that have not developed resistance, the pesticide resistance of pests and diseases increases with the number of times the chemical agent is used in the monitored area. Therefore, in this embodiment, the monitoring requirements for the pesticide resistance can be the cumulative resistance after multiple uses of the chemical agent, or the increased resistance relative to the last use of the chemical agent, or other monitoring requirements related to pesticide resistance defined by plant protection experts, so the calculation method of resistance varies with different monitoring results.

[0071] The first sample plot is used as the sampling result of the tobacco field, and the pesticide resistance of the first sample plot can be directly used to reflect the pesticide resistance of the tobacco field. In order to further improve the accuracy of the pesticide resistance of the tobacco field, the pesticide resistance result of the first sample plot can also be adjusted. For example, the weight can be set according to the change ratio of the standard deviation of the previously recorded pesticide resistance data, or according to the proportion of pest and disease populations in all the plots of the tobacco field; or the pesticide resistance of multiple first samples is used for comprehensive calculation, which can be the average or median of multiple first samples.

[0072] In this embodiment, the resistance to pesticides is indirectly quantified by the change in the activity of the pests and diseases after the use of chemical agents. As a preferred technical solution of this application, the activity of the monitored area without the use of any chemical agents is obtained, recorded as the initial activity; the activity of the monitored area after the first use of chemical agents is obtained, recorded as the first activity; the activity of the monitored area after the use of chemical agents for the second time is obtained, recorded as the second activity; the first difference between the second activity and the first activity is calculated, and the ratio of the first difference to the initial activity is used as the resistance of the monitored area.

[0073] When chemical agents are used in the initial stage, the number of pests and diseases that develop resistance is relatively small. In this embodiment, this part of the pests and diseases is chosen to be ignored. Therefore, the activity of the pests and diseases that survive the first use of chemical agents in the monitored area is taken as the activity without the influence of resistance, which is recorded as the first activity. The activity of the monitored area after the second and subsequent multiple uses of chemical agents is recorded as the second activity. After the monitored area uses chemical agents multiple times, the resistance of the pests and diseases in the area increases, and the activity of the pests and diseases increases relative to the activity of the pests and diseases after the first use of the agent, that is, the increased activity comes from the selective effect of resistance on the pests and diseases. Therefore, in this embodiment, the change of the second activity relative to the first activity is calculated first, and then the proportion of the change to the initial activity is calculated, and the proportion of the activity in the initial activity that is affected by resistance is used to indirectly quantify the resistance.

[0074] In order to solve the problem that it is difficult to accurately quantify the resistance of pests and diseases after using chemical agents, the movement position of pests and diseases in the first sample is obtained by image recognition, and then the activity of pests and diseases is calculated by the movement position. The resistance is quantified by the ratio of the change in activity caused by resistance to the activity of pests and diseases without using chemical agents. This is helpful to observe the effect of chemical agents after using chemical agents in the first sample, so as to adjust the subsequent pest and disease control measures in a targeted manner.

[0075] The size of the active area is less than or equal to the size of the area to be monitored, and is affected by the body size of the pests and the movement mode. If the proportion of the body size of the pests to the horizontal area of ​​the area to be monitored is greater than or equal to the body-area ratio threshold, the body size of the pests is also included in the active area and the movement area is calculated, such as the second pest; if the proportion of the body size of the pests to the horizontal area of ​​the area to be monitored is less than the body-area ratio threshold, the body size of a single pest can be ignored, such as the first pest, the third pest, and the fourth pest, and the pests are regarded as points in the active area, and then the size of the active area is calculated.

[0076] As a preferred technical solution of the present application, the method for determining the active area of ​​the pests and diseases based on the multiple movement positions includes: presetting a size-area ratio threshold, calculating a first ratio of the size of the pests and diseases to the horizontal area of ​​the area to be monitored, comparing the first ratio with the size-area ratio threshold, and determining a method for calculating the active area based on the comparison result and the monitoring category.

[0077] For the first pest, due to its wide active area, it can be regarded as a point relative to the first plot, and its actual movement trajectory is curved, and the overlap of the movement trajectory is low. At the same time, the area of ​​the first plot is much larger than the size of the first pest, and relative to the movement trajectory, there are more useless spaces in the first plot where pests will not fly, and the linear characteristics of the active area of ​​the first pest are more obvious. Therefore, in this embodiment, preferably, the top view picture of the first plot is used to calculate the movement distance of the first pest on the horizontal plane.

[0078] The preferred calculation method of the active area is: when reaching multiple time steps, obtain multiple images of the first pest to be monitored, and fuse the first pest in the multiple images to be detected into one image through an image fusion algorithm. Extract the contour features of the first pest, obtain the first number of the first pests with the same contour features, and remove the first pests whose first number is less than the preset first threshold or whose continuous missing number is greater than the preset first threshold. Connect the movement positions of the first pests remaining in the fused image in sequence according to the order of the time steps, and obtain the movement distance of the first pest by coordinate calculation.

[0079] Since pests and diseases that are larger in size and capable of flying have more obvious linear characteristics in their active areas in the sample plot, calculating the movement distance of pests and diseases can more accurately reflect their active areas, which helps to reduce the impact of useless space areas in the sample plot on the calculated activity.

[0080] For the second pest, due to its large size, the movement trajectory also occupies a certain area, and the active area is the single plant, and the overlap of the movement trajectory under multiple time steps is high. At the same time, the second pest mainly moves on the leaves of the single plant. In the space occupied by the single plant, it will not or is difficult to involve the useless space in the vertical direction between the leaves, and the planar characteristics of its active area are more obvious. Therefore, in this embodiment, preferably, the top view picture of the single plant is used to calculate the movement area of ​​the second pest in the horizontal plane. The preferred calculation method of the active area is: when reaching multiple time steps, obtain multiple pictures of the second pest to be monitored of a single or multiple pests, fuse the graphics of the second pest in the multiple pictures to be detected into one picture through the image fusion algorithm, perform edge detection on the fused picture, and obtain the fused contour feature. Connect the fused contour features close to the edge of the fused picture, and the area of ​​the figure surrounded by the fused contour features after the connection is closed is the movement area.

[0081] Since pests and diseases that are larger in size and capable of crawling have more obvious planar characteristics of their active areas in the sample plots, calculating the movement area of ​​pests and diseases can more accurately reflect their active areas on the leaves, which helps to reduce the impact of useless space areas in the sample plots, especially in the vertical direction, on the calculated activity. At the same time, by connecting the contour features of pests and diseases close to the edge of the image, the body shape of the pests and diseases themselves is included in the figure surrounded by the fused contour features, thereby improving the accuracy of calculating the movement area.

[0082] For the third pest, since its individual active area is small, it can be regarded as a point relative to the single plant area, and its actual movement trajectory can be regarded as a curve. However, the third pest is a group pest, and its number is large in the single plant area. Therefore, the movement trajectories of different individuals of the third pest have a high degree of overlap, which makes it difficult to track individuals, and they are easy to interfere with each other, and the spatial characteristics of their activity areas are more obvious. Therefore, in this embodiment, preferably, the top view and side view of the single plant area are used to calculate the movement space of the third pest in space.

[0083] The preferred calculation method of the active area is: when reaching multiple time steps, obtain multiple top-view pictures and multiple side-view pictures of the single plant area. The multiple top-view pictures are fused, and the multiple third pests and diseases in the fused pictures form a first point set. The convex hull algorithm is used to draw polygons on the first point set to obtain the horizontal active area of ​​the third pests and diseases. The multiple side-view pictures are fused, and the multiple third pests and diseases in the fused pictures form a second point set. The vertical distance between the highest and lowest positions of the third pests and diseases in the second point set is calculated to obtain the maximum vertical movement height of the third pests and diseases. The product of the area of ​​the horizontal active area of ​​the third pests and diseases and the maximum vertical movement height is calculated to obtain the size of the movement space of the third pests and diseases.

[0084] Since pests and diseases that are small in size and capable of flying appear in sample plots in the form of group aggregation, the spatial characteristics of their active areas are more obvious. By calculating the volume of the movement space where pests and diseases gather, the active areas of their groups can be more accurately reflected, which helps to reduce the difficulty of image recognition and reduce repeated active area calculations caused by mutual influence between pests and diseases individuals.

[0085] The fourth pests and diseases include pests and diseases that move as a single individual or as a group. Due to the limited movement ability of a single individual, the active area is the leaves of the single plant, and the planar characteristics of the active area are more obvious. Therefore, in this embodiment, preferably, a top-down image of the leaves of the single plant is used to calculate the active area of ​​the fourth pest and disease. The preferred calculation method of the active area is: when multiple time steps are reached, multiple top-down images of the leaves are obtained. The multiple top-down images of the leaves are fused, and the multiple fourth pests and diseases in the fused images form a third point set. The convex hull algorithm is used to draw a polygon for the third point set, and the polygon is the fused contour feature of the fourth pest and disease. The area enclosed by the fused contour feature of the fourth pest and disease is calculated to obtain the movement area of ​​the fourth pest and disease.

[0086] Embodiment 2: The difference between this embodiment and embodiment 1 is that: a system, device and storage medium for implementing a pest detection method are provided.

[0087] A tobacco field pest and disease monitoring system based on image recognition, e.g. Figure 2 As shown, the method for performing the tobacco field pest and disease monitoring method based on image recognition comprises:

[0088] A classification module is used to divide the tobacco field into sample plots; after selecting the first sample plot, an initial image of the first sample plot is received, the initial image is output to a pest classification model, the appearance features of the pests are extracted for classification, and a monitoring category of the pests is output according to the type of the pests and the appearance features.

[0089] The monitoring module is used to receive the appearance features and the monitoring category, adjust the area to be monitored for the pests and diseases, re-photograph the area to be monitored, and output the image to be monitored.

[0090] The activity module is used to receive the movement position of the pest in the image to be monitored when the preset time step is reached, determine the active area of ​​the pest based on multiple movement positions, and estimate and output the pesticide resistance of the tobacco field according to the pesticide resistance of the first sample plot. The activity module includes an image fusion submodule, a plane drawing submodule and a calculation submodule.

[0091] The pesticide resistance module is used to receive the activity recorded in the history and the current activity and output the pesticide resistance of the first sample plot; and estimate and output the pesticide resistance of the tobacco field according to the value of the pesticide resistance.

[0092] The image fusion submodule is used to receive the pests and diseases in multiple images to be monitored, fuse all the pests and diseases in different movement positions into one image, and output the fused image.

[0093] The plane drawing submodule is used to receive the fused image, extract different movement positions of the pests in the fused image, and draw the plane figure formed by the pests based on the different movement positions. The plane drawing submodule includes an edge detection submodule and a convex hull algorithm submodule.

[0094] The calculation submodule is used to calculate the geometric size of the active area, and calculate the activity based on the geometric size and a preset reference value.

[0095] The edge detection submodule is used to obtain the contour features of the pests and diseases through an edge detection algorithm, and after connecting all the contour features, draw a plane figure formed by the pests and diseases.

[0096] The convex hull algorithm submodule is used to receive the movement positions of all the pests and diseases in the fused image, extract the coordinates, and use the convex hull algorithm on all the movement positions to draw a plane figure formed by the pests and diseases.

[0097] A tobacco field pest and disease monitoring device based on image recognition, the device comprising: an image or video sensor for collecting tobacco field and pest and disease images; at least one processor capable of processing the image data collected by the sensor; at least one memory for storing tobacco field and pest and disease image data collected by the sensor and at least one program; when the at least one program is executed by the at least one processor, the at least one processor implements the tobacco field pest and disease monitoring method based on image recognition, the processor and the memory can be connected by a bus or other means for storing at least one program; the memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to the device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof, when the at least one program is executed by the at least one processor, the at least one processor implements the tobacco field pest and disease monitoring method based on image recognition.

[0098] A computer storage medium stores a program executable by a processor, wherein the program executable by the processor is used to implement the tobacco field pest and disease monitoring method based on image recognition when executed by the processor.

[0099] It will be appreciated by those skilled in the art that all or some of the steps and systems in the disclosed method above may be implemented as software, firmware, hardware and appropriate combinations thereof. Some physical components or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor or a microprocessor, or may be implemented as hardware, or may be implemented as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or a non-transitory medium) and a communication medium (or a temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that may be used to store desired information and may be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0100] The above embodiments are only used to illustrate the present invention and are not intended to limit the technical solutions described in the present invention. Although the present invention has been described in detail with reference to the above embodiments, the present invention is not limited to the above specific implementation methods. Therefore, any modification or equivalent replacement of the present invention; and all technical solutions and improvements thereof that do not depart from the spirit and scope of the invention are included in the scope of the claims of the present invention.

Claims

1. A tobacco field pest and disease monitoring method based on image recognition, characterized in that the steps include: Sampling the tobacco field to obtain a first sample; Obtaining an initial image of the first sample plot, inputting the initial image into a pest classification model, classifying the pests according to their appearance characteristics, and dividing the pests into monitoring categories according to their types and appearance characteristics; Re-determine the area to be monitored for the pests and diseases according to the appearance characteristics and the monitoring category, and re-photograph the area to be monitored to obtain a picture to be monitored; When the preset time step is reached, different movement positions of the pests and diseases in the image to be monitored are obtained, and the active area of ​​the pests and diseases is determined based on the multiple movement positions; Calculating the activity of the pests and diseases according to the geometric size of the active area; Calculate the drug resistance of the first sample plot according to the activity recorded in the history and the current activity; Using the pesticide resistance of the first sample plot to estimate the pesticide resistance of the tobacco field; The method of classifying the pests and diseases into monitoring categories according to the types of pests and diseases and the appearance characteristics thereof includes: Determine the size of the pests and diseases according to the appearance characteristics; determine the movement mode of the pests and diseases according to the type of the pests and diseases and the appearance characteristics; preset a body size threshold; compare the body size of the pests and diseases with the body size threshold, and classify the pests and diseases into monitoring categories according to the comparison result and the movement mode; Determining the movement mode of the pests and diseases according to the types of the pests and diseases and the appearance characteristics includes: The appearance features include contour features and movement organ features, and the movement organ features include flight organ features and foot organ features; the movement mode of the pests is determined based on the type of the pests and the movement organ features.

2. The tobacco field pest and disease monitoring method based on image recognition according to claim 1, characterized in that: The step of calculating the activity of the pests and diseases according to the geometric size of the active area includes: Obtain active areas of multiple different clusters of the pests and diseases when the pests and diseases are not affected by any chemical agents, calculate the geometric sizes of the active areas, and determine a benchmark value based on the multiple geometric sizes; take the ratio of the geometric size of the active area at that time to the benchmark value as the activity level of the pests and diseases.

3. The tobacco field pest and disease monitoring method based on image recognition according to claim 1, characterized in that: The calculating the drug resistance of the first sample according to the activity recorded in the historical records and the current activity includes: Obtain the activity of the monitored area before using any chemical agent, recorded as the initial activity; obtain the activity of the monitored area after using chemical agents for the first time, recorded as the first activity; obtain the activity of the monitored area after using chemical agents for the second time, recorded as the second activity; calculate the first difference between the second activity and the first activity, and take the ratio of the first difference to the initial activity as the drug resistance of the monitored area.

4. A tobacco field pest and disease monitoring method based on image recognition according to any one of claims 1 to 3, characterized in that: A size-area ratio threshold is preset, a first ratio of the size of the pests and diseases to the horizontal area of ​​the monitored area is calculated, the first ratio is compared with the size-area ratio threshold, and a calculation method of the active area is determined according to the comparison result and the monitoring category.

5. The tobacco field pest and disease monitoring method based on image recognition according to claim 4 is characterized in that: The calculation method of the active area includes: when reaching multiple time steps, obtaining the adjacent movement positions of a single or multiple pests and diseases, calculating the coordinate distance according to the coordinates of the movement position, and summing the multiple coordinate distances to obtain the cumulative movement distance.

6. The tobacco field pest and disease monitoring method based on image recognition according to claim 4, characterized in that: The calculation method of the active area includes: when reaching multiple time steps, obtaining multiple images of the to-be-monitored single or multiple pests and diseases, fusing the contour features of the pests and diseases in the multiple images to be monitored into one image through an image fusion algorithm, extracting the contour features of the pests and diseases in the fused image, and obtaining fused contour features; calculating the area of ​​the figure enclosed by the fused contour features to obtain the motion area.

7. The method for monitoring tobacco field pests and diseases based on image recognition according to claim 4, characterized in that: The calculation method of the active area includes: When a plurality of the time steps are reached, a plurality of top-view images and a plurality of side-view images of the area to be monitored are acquired; Fusing the plurality of overhead images, drawing polygons for the first point set formed by the pests and diseases in the fused images, and obtaining a horizontal active area; Fusing the multiple side-view images to obtain a second point set formed by the pests and diseases in the fused image, calculating the vertical distance between the pests and diseases at the highest position and the pests and diseases at the lowest position in the second point set, and obtaining the maximum vertical movement height; The size of the motion space is obtained by calculating the product of the area of ​​the horizontal active region and the maximum vertical motion height.

8. A tobacco field pest and disease monitoring system based on image recognition, characterized in that: The classification module is used to divide the tobacco field into plots; after selecting the first plot, receive the initial image of the first plot, output the initial image to the pest classification model, extract the appearance features of the pests for classification; output the monitoring category of the pests according to the type of pests and the appearance features, including: Determine the size of the pests and diseases according to the appearance features; the appearance features include contour features and movement organ features, and the movement organ features include flight organ features and foot organ features; determine the movement mode of the pests and diseases according to the type of the pests and diseases and the movement organ features; preset a body size threshold; compare the body size of the pests and diseases with the body size threshold, and classify the pests and diseases into monitoring categories according to the comparison result and the movement mode; A monitoring module, for receiving the appearance features and the monitoring category, adjusting the area to be monitored for the pests and diseases; re-photographing the area to be monitored, and outputting a picture to be monitored; The activity module is used to receive the movement position of the pest in the image to be monitored when the preset time step is reached, determine the active area of ​​the pest based on multiple movement positions; calculate and output the activity of the pest according to the geometric size of the active area; The pesticide resistance module is used to receive the activity recorded in the history and the current activity and output the pesticide resistance of the first sample plot; and estimate and output the pesticide resistance of the tobacco field according to the pesticide resistance of the first sample plot.

Citation Information

Patent Citations

  • Method for rapidly identifying drug resistance of spodoptera frugiperda to bisamide insecticides

    CN112056277A

  • Automatic monitoring and intelligent drug delivery prevention and control system for crop pest in facility greenhouse

    CN105137840A

  • Crop disease and insect pest type identification method, system and device, and medium

    CN111814622A

  • Crop plant disease and insect pest plant protection reminding method and device, electronic equipment and storage medium

    CN113657771A

  • Interaction method and device in virtual reality scene, equipment and storage medium

    CN116048281A