Silkworm disease prevention and control method and system for predicting pathogen trend based on artificial intelligence

Abstract

The invention relates to the technical field of silkworm disease prevention and control, in particular to a silkworm disease prevention and control method and system for predicting pathogen trend based on artificial intelligence. The method comprises the following steps: screening pathogen monitoring sites from silkworm industrial sites, and collecting samples to be detected from the pathogen monitoring sites based on a preset sampling rule; a detection result and corresponding weather data during pathogen detection are stored in a pathogen database; setting a to-be-predicted pathogen type and initial prediction time, and starting from the initial prediction time, obtaining weather forecast data of multiple days in the future; inputting the weather forecast data into the pathogen prediction model to obtain pathogen concentrations of pathogen types in multiple days in the future, and recording the pathogen concentrations as a first prediction content set; and grading the first predicted content set, and determining early warning and prevention and control measures according to a grading result. A pathogen prediction model based on artificial intelligence is used for receiving weather forecast data, the proliferation result of pathogens in multiple days in the future is predicted, prevention and control measures can be laid out in advance on silkworm industry points, and the risk of pathogen outbreak is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of silkworm disease prevention and control, and in particular to a silkworm disease prevention and control method and system based on artificial intelligence prediction of pathogen trends. Background Art

[0002] Silkworm is a traditional industry in my country and a dominant industry for accelerating the development of high-quality and efficient modern characteristic agriculture. In the breeding process, silkworm disease is one of the main reasons for the loss of economic benefits in silkworm breeding. In the process of silkworm breeding, many individual breeders did not conduct scientific and thorough disinfection in the early stage, which made it easy for pathogens to accumulate, leading to large-scale outbreaks of silkworm disease. After the outbreak, due to unclear disinfection priorities, repeated disinfection, missed disinfection, and wrong disinfection occurred, causing secondary outbreaks. In very few silkworm areas, a single season of harvest was completely lost, and even silkworms could not be raised. In order to improve the quality and benefits of silkworm breeding, it is necessary to detect, discover, respond and treat silkworm disease pathogens early.

[0003] In the prior art, visual diagnosis is one of the methods for farmers to judge whether silkworms are sick. It is to directly observe and preliminarily evaluate whether the physiological activities of silkworms are abnormal. Although visual diagnosis can provide some intuitive information, it is difficult to find pathogenic microorganisms to silkworms because it relies on naked eye observation, and its accuracy and reliability are relatively low. Microscopic examination is to use tools such as microscopes to observe the collected pathogen samples more carefully. The samples can be magnified so that the inspectors can observe tiny pathogens. Although this method is more accurate than visual diagnosis, its sensitivity is limited by the resolution of the microscope and the experience of the observer. It is difficult to accurately determine the type of pathogen. For pathogen samples with low concentrations, microscopic examination is also difficult to accurately detect. In recent years, polymerase chain reaction (PCR) fluorescence quantitative detection technology, as a molecular biological detection method, can amplify the genetic information of pathogenic microorganisms through the amplification of pathogenic microorganism DNA and RNA characteristic sequences, and quickly and efficiently detect the types of pathogenic microorganisms. With the help of pre-prepared gene probes, qualitative and quantitative detection of pathogens can be achieved.

[0004] However, if these detection methods are only used for confirmatory detection after silkworms become ill, not only will the timing of applying prevention and control measures be delayed, but also the need to use drugs with higher concentrations during the treatment of silkworms will increase the probability of pathogens developing drug resistance, and it will also be difficult to recover the economic losses that have already been caused. At present, there are situations in the agricultural field where artificial intelligence is used to guide agricultural production. By collecting a large amount of historical data through big data, artificial intelligence can analyze and learn more valuable information and features, and predict the future development trends of these data. Therefore, how to use artificial intelligence technology to make up for the lag of existing detection methods is a problem that needs to be solved urgently. This application proposes a silkworm disease prevention and control method based on artificial intelligence to predict pathogen trends. According to the sample test results of the early stage of the development of silkworm disease pathogens, the future development trend and risk of the pathogen are estimated, and the current silkworm breeding environment is evaluated and predicted, so as to deploy prevention and control measures in a targeted manner in advance, and guide the precise disinfection of silkworm rooms, silkworm tools, and surrounding environments. Summary of the invention

[0005] The purpose of the present invention is to: Due to the lag in the diagnosis and detection of silkworm pathogens in the prior art, the prevention and control measures are not arranged in time. In response to this problem, this application provides a silkworm disease prevention and control method and system based on artificial intelligence to predict pathogen trends, and builds a pathogen prediction model based on the long short-term memory neural network LSTM. With the help of the slow execution characteristics of LSTM, the characteristics of weather data affecting the pathogen proliferation process in the breeding cycle are learned. Then, based on the weather forecast data for many days in the future, the proliferation results of pathogens in the next few days are predicted, which is conducive to the early layout of prevention and control measures at the silkworm industry points and reduces the risk of pathogen outbreaks.

[0006] In order to achieve the above-mentioned object of the invention, the present invention provides the following technical solutions: Silkworm disease prevention and control methods based on artificial intelligence prediction of pathogen trends include: Select pathogen monitoring points from silkworm industry points, collect samples to be tested from the pathogen monitoring points based on preset sampling rules, and test the samples to be tested; the sampling rules include sampling timing rules and sampling fixed point rules; The obtained detection results and the weather data corresponding to the pathogen detection are stored in the pathogen database, and the detection results are fed back to the pathogen monitoring point; the type of the pathogen to be predicted and the starting prediction time are set, and the weather forecast data for the next multiple days are obtained starting from the starting prediction time; The weather forecast data is input into the pathogen prediction model to obtain the pathogen concentration of the pathogen type in the future multiple days, which is recorded as a first predicted content set; the first predicted content set is classified, and early warning and prevention and control measures are determined according to the classification results.

[0007] As a preferred technical solution of the present application, the grading of the first predicted content set includes: The pathogen concentration is graded, and the pathogen concentration corresponding to the first predicted content set is converted into a first graded set; the number of each grade in the first graded set is counted, and the severity of the pathogen in the first graded set is determined according to the number of the grades.

[0008] As a preferred technical solution of the present application, before grading the first grading set, the method further includes smoothing the first grading set, and the method includes: Preset a first threshold and a second threshold; obtain a first subset whose number of grades is greater than the first threshold, and the grades in the first subset are the same and continuous in time; obtain a second subset adjacent to the first subset, if the number of continuous grades in the second subset is less than or equal to the second threshold, regard the second subset as a discrete data set, and modify the second subset so that its grade is the same as that of the first subset.

[0009] As a preferred technical solution of the present application, the grading of the pathogen concentration includes: A first concentration threshold and a second concentration threshold are preset; the pathogen concentration is classified as low concentration if it is less than or equal to the first concentration threshold, as medium concentration if it is greater than the first concentration threshold and less than or equal to the second concentration threshold, and as high concentration if it is greater than the second concentration threshold.

[0010] As a preferred technical solution of the present application, the method for screening pathogen monitoring points from silkworm industry points includes: Obtain the output value and the incidence rate of silkworm diseases of multiple silkworm industry sites in the previous year, calculate the ratio of the incidence rate to the output value, and obtain the first loss rate of the silkworm industry sites caused by silkworm diseases; sort the first loss rates, and obtain multiple silkworm industry sites with top rankings as the pathogen monitoring sites.

[0011] As a preferred technical solution of the present application, after determining the pathogen monitoring point, it also includes: Obtain the number of roads and traffic flow between the pathogen monitoring point and other silkworm industry points; determine the infection weight of the pathogen monitoring point to other silkworm industry points based on the number of roads and the traffic flow; and add the silkworm industry points whose infection weight is greater than a preset infection threshold to the pathogen monitoring point.

[0012] As a preferred technical solution of the present application, it includes: obtaining weather data of multiple silkworm industry locations, dividing the silkworm industry locations into multiple sets based on the weather data; and conducting joint early warning and prevention and control for the silkworm industry locations in the same set.

[0013] As a preferred technical solution of the present application, the method of dividing the silkworm industry sites into multiple sets based on the weather data includes: Acquire the historical temperature and humidity of the plurality of silkworm industry sites, perform cluster analysis on the historical temperature and humidity of each silkworm industry site to obtain a plurality of temperature and humidity data sets, and encode the temperature and humidity data sets into a first clustering vector; The similarity between the silkworm industry sites is calculated based on the first clustering vector, recorded as the first similarity, and the silkworm industry sites whose first similarity satisfies a preset error range are divided into the same set.

[0014] As a preferred technical solution of the present application, the method of dividing the silkworm industry sites into multiple sets based on the weather data includes: Obtaining historical wind directions and wind forces of a plurality of said silkworm industry points, counting the number of days with wind direction, and combining the number of days into a wind direction vector; counting the number of days with wind force, and combining the number of days into a wind force vector; directly combining the wind direction vector and the wind vector, and encoding them into a second clustering vector; The similarity between the silkworm industry sites is calculated based on the second clustering vector, recorded as the second similarity, and the silkworm industry sites whose second similarity satisfies a preset error range are divided into the same set.

[0015] This application also provides a silkworm disease prevention and control system based on artificial intelligence to predict pathogen trends, including: A sampling module is used to select pathogen monitoring points from silkworm industry points, collect samples to be detected from the pathogen monitoring points based on preset sampling rules, and detect the samples to be detected; the sampling rules include sampling timing rules and sampling fixed point rules; An acquisition module is used to store the obtained detection results and the weather data corresponding to the pathogen detection in the pathogen database, and to feed back the detection results to the pathogen monitoring point; set the pathogen type to be predicted and the starting prediction time, and obtain the weather forecast data for the next several days starting from the starting prediction time; The prediction module inputs the weather forecast data into the pathogen prediction model to obtain the pathogen concentration of the pathogen type in the future multiple days, which is recorded as a first prediction content set; the first prediction content set is classified, and early warning and prevention and control measures are determined according to the classification results.

[0016] Compared with the prior art, the present invention has the following beneficial effects: 1. In view of the problem of qualitative and quantitative detection of the proliferation trend of pathogenic microorganisms in the silkworm breeding environment in various periods, scenes and links of silkworm industry production, a pathogen prediction model is introduced to evaluate and predict the current disease risk of the silkworm breeding environment, guide the precise disinfection of silkworm rooms, silkworm tools and surrounding environments, and provide a scientific basis for producers to take timely prevention and control measures before the outbreak of silkworm diseases. Effectively improving the quality and benefits of silkworm breeding provides important support for increasing silkworm output value and industrial development, and is conducive to promoting industry stability.

[0017] 2. The classification of silkworm industry sites is different, and the specific area size is different. The smaller the area, the more sites and data need to be analyzed and tested, and the more accurate the trend prediction results of the model; the larger the area, the impact of regional commonalities on the development trend of pathogens can be analyzed, for example, the implementation of prevention and control policies in administrative regions, the undulating trend of terrain and the development trend of pathogen infection are consistent, which helps to fully grasp the distribution of pathogens.

[0018] 3. The environmental temperature and humidity are different in different periods, and the ability of pathogen proliferation is also different. There is a correlation between the number of pathogens and weather data. Therefore, collecting samples to be tested at different periods and times for analysis is conducive to modeling the process of weather affecting pathogen proliferation, so that the pathogen prediction model can learn the characteristics of this impact process and improve the accuracy of the prediction. Silkworm tools, mulberry leaves, and silkworm feces collection tools have the highest correlation with silkworm activities and the highest possibility of detecting pathogens. At the same time, pathogens and other microorganisms are easily infected to other places through the air. Sampling pathogens in multiple locations such as the surrounding environment of the silkworm room can not only increase the number of samples to be tested, but also take targeted prevention and control measures based on the prediction results of different locations, thereby reducing the cost of prevention and control. The sampling timing rules are formulated according to the time law of pathogen accumulation in the silkworm breeding cycle, and the sampling fixed point rules are formulated according to the regional law of the appearance of historical positive samples, which improves the scientificity and pertinence of sampling, thereby improving the accuracy and efficiency of pathogen detection.

[0019] 4. The output value of silkworm industry sites can be used to evaluate their breeding scale, business scale and economic benefits. The higher the output value, the larger the breeding scale and the more frequent the business activities. If a silkworm disease outbreak occurs, the greater the loss. Therefore, it is representative and effective to use it as a pathogen monitoring site. At the same time, more samples to be tested can be collected, which improves the accuracy of the pathogen prediction model.

[0020] 5. By constructing a nonlinear regression function for the number of roads and traffic flow at the pathogen monitoring points, the product of the coefficients of the two variables is calculated. This can not only be used to analyze the degree of influence of pathogen infection caused by traffic between silkworm industry points, but also to analyze the influence coefficient of the number of roads and traffic flow on pathogens, which can provide a basis for the selection of silkworm industry points.

[0021] 6. Weather is one of the important factors affecting pathogen growth during the silkworm breeding cycle. Among the silkworm industry sites with similar weather conditions, the pathogen growth process is also similar. By identifying these similar silkworm industry sites based on weather data, unified prevention and control measures can be taken for these sites. There is no need to conduct pathogen detection at each site, which helps improve the efficiency of pathogen monitoring and helps other undetected sites to quickly respond to prevention and control.

[0022] 7. Weather has a lasting impact on the proliferation of pathogens, and the pathogen proliferation process itself also has a quantitative relationship, which makes the time series of pathogen concentration have a contextual relationship. In order to capture this contextual relationship, it is preferred to build a pathogen prediction model based on the LSTM long short-term memory neural network, and retain the contextual features of the weather data affecting the pathogen proliferation process through cell units, so that the prediction results also have a persistent relationship, which is conducive to accurately reflecting the development trend of pathogen concentration. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a flow chart of the silkworm disease prevention and control method; Figure 2 A schematic diagram of the process for determining pathogen monitoring points; Figure 3 A schematic diagram of the process of clustering silkworm industry sites based on temperature and humidity; Figure 4 It is a schematic diagram of the process of clustering the silkworm industry sites based on wind force and wind direction; Figure 5 This is a structural diagram of the silkworm disease prevention and control system. DETAILED DESCRIPTION

[0024] 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.

[0025] 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.

[0026] 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.

[0027] 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.

[0028] 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.

[0029] Example 1: See Figure 1-Figure 5 As shown, the silkworm disease prevention and control method based on artificial intelligence prediction of pathogen trends provided in this embodiment includes: See also Figure 1 As shown, pathogen monitoring points are selected from silkworm industry points, and the silkworm industry points can be silkworm breeding points; they can also be classified based on administrative regions, including villages, townships, counties, etc. The classification of silkworm industry points is different, and the specific area size is different. The smaller the area, the more points and data need to be analyzed and tested, and the more accurate the trend prediction results of the model; the larger the area, the impact of regional commonalities on the development trend of pathogens can be analyzed, for example, whether the implementation of prevention and control policies in administrative regions, the undulating trend of terrain and the development trend of pathogen infection are consistent, which helps to fully grasp the distribution of pathogens.

[0030] Before the start of each annual silkworm season, at different silkworm ages, and after the end of the silkworm season, samples to be tested are collected from the pathogen monitoring points based on preset sampling rules, and the samples to be tested are tested. The sampling rules include sampling timing rules and sampling fixed-point rules. According to the time law of pathogen accumulation during the silkworm breeding cycle, sampling timing rules are formulated. For example, the main growth stage of silkworms is the insect stage, with weak resistance, and it is also the main period for the occurrence and development of silkworm diseases; while the egg stage and pupa stage have corresponding shell protection; the moth stage has the ability to fly, is less affected by external pathogens, and is difficult to get sick. Therefore, each insect age stage in the insect stage needs to be tested, and the number of tests in other periods is reduced. Temperature will affect the activity of pathogens, so three sampling time points in the morning, middle and evening are determined according to the daily temperature difference. According to the regional law of the appearance of historical positive samples during the silkworm breeding process, such as in the silkworm room, sampling fixed-point rules are formulated. For example, silkworm tools, mulberry leaves, and silkworm feces collection tools.

[0031] The environmental temperature and humidity are different in different periods, and the ability of pathogen proliferation is also different. There is a correlation between the number of pathogens and weather data. Therefore, collecting samples to be tested at different periods and times for analysis is conducive to modeling the process of weather affecting pathogen proliferation, so that the pathogen prediction model can learn the characteristics of this impact process and improve the accuracy of the prediction. Silkworm tools, mulberry leaves, and silkworm feces collection tools have the highest correlation with silkworm activities and the highest possibility of detecting pathogens. At the same time, pathogens and other microorganisms are easily infected with air in other places. Sampling pathogens in multiple locations such as the surrounding environment of the silkworm room can not only increase the number of samples to be tested, but also take targeted prevention and control measures based on the prediction results of different locations, thereby reducing the cost of prevention and control. The sampling timing rules are formulated according to the time law of pathogen accumulation in the silkworm breeding cycle, and the sampling fixed point rules are formulated according to the regional law of the appearance of historical positive samples, which improves the scientificity and pertinence of sampling, thereby improving the accuracy and efficiency of pathogen detection.

[0032] In this embodiment, the detection methods that can be used include microscopic examination, specific antigen screening, nucleic acid amplification detection technology, such as fluorescent quantitative PCR detection, etc. If the test result of the sample to be detected is positive, it means that the silkworm disease pathogen is contained. Further, the obtained test results, such as pathogen type, nucleic acid amplification cycle number, pathogen concentration, and weather data corresponding to the pathogen detection, are stored in the pathogen database, and the test results are fed back to the pathogen monitoring point.

[0033] Select the pathogen type to be predicted; obtain multiple pathogen concentrations of the pathogen type from the pathogen database in chronological order; and obtain historical and future weather forecast data for multiple days starting from a preset starting prediction time.

[0034] The time series of the pathogen concentration and the corresponding time series of the historical weather data are first input into the pathogen prediction model, so that the pathogen prediction model is first calibrated based on the pathogens and weather data that have been detected in the current silkworm breeding cycle. Then the weather forecast data is input into the pathogen prediction model to obtain the pathogen concentration of the pathogen type in the future multiple days, which is recorded as the first predicted content set. The first predicted content set is graded, and early warning and prevention and control measures are determined according to the classification results.

[0035] In view of the need to qualitatively and quantitatively detect the proliferation trend of pathogenic microorganisms in the silkworm breeding environment at all stages, scenes and links of silkworm production, a pathogen prediction model is introduced to evaluate and predict the current risk of disease in the silkworm breeding environment, guide the precise disinfection of silkworm rooms, silkworm tools and surrounding environments, and provide a scientific basis for producers to take timely prevention and control measures before the outbreak of silkworm diseases. Effectively improving the quality and benefits of silkworm breeding provides important support for increasing silkworm output value and industrial development, and is conducive to promoting industry stability.

[0036] As a preferred technical solution of the present application, the grading of the first predicted content set includes: The pathogen concentration is graded, and the pathogen concentration corresponding to the first predicted content set is converted into a first graded set. The number of each grade in the first graded set is counted, and the pathogen severity of the first graded set is determined according to the number of the grades. For example, the grades are set to low, medium, and high, and the ratio of each grade to the total number of the first predicted content set is calculated. If the ratio corresponding to the low is more than 80%, the pathogen severity is not serious; if the ratio corresponding to the medium is 30%-50%, and the ratio corresponding to the high is less than 20%, the pathogen severity is relatively serious; if the ratio corresponding to the high is more than 30%, the pathogen severity is very serious. Optionally, weight scores are set for different grades, for example, 1 point, 2 points, and 3 points are set for low, medium, and high, and the product is calculated with the number of corresponding grades. The weighted product results are summed, and the ratio of the summation result to the total number of grades is calculated to obtain the weighted average score of the first graded set. The pathogen severity of the first predicted content set is determined by comparing the weighted average score with the threshold set by the expert.

[0037] Furthermore, the first classification set is smoothed. Preferably, a first threshold and a second threshold are preset. A first subset whose number of classifications is greater than the first threshold is obtained, and the classifications in the first subset are the same and continuous in time; a second subset adjacent to the first subset is obtained, and if the number of continuous data in the second subset is less than or equal to the second threshold, the second subset is regarded as a discrete data set, and the second subset is modified so that its classification is the same as that of the first subset. The smoothing step is performed repeatedly until the number of discrete data sets is 0. It is difficult to achieve completely accurate predicted pathogen concentration content. If there are discrete data points, it is easy to cause interference when evaluating the development trend of pathogens. In addition, the change in pathogen concentration has continuity and contextual relationship. In a natural environment, it is difficult for discrete values ​​that deviate too much to appear. Smoothing the first classification set helps to analyze the continuity of the development trend of pathogens.

[0038] Furthermore, the grading of the pathogen concentration includes: presetting a first concentration threshold and a second concentration threshold; the pathogen concentration is classified as low concentration when it is less than or equal to the first concentration threshold, as medium concentration when it is greater than the first concentration threshold and less than or equal to the second concentration threshold, and as high concentration when it is greater than the second concentration threshold. Grading the pathogen concentration helps to intuitively reflect the severity of the pathogen development trend, and when smoothing the first grading set, it helps to simplify the data analysis process.

[0039] Example 2: See Figure 2As shown, this embodiment further refines the technical features of screening the pathogen monitoring points on the basis of the first embodiment.

[0040] As a preferred technical solution of the present application, the method for selecting pathogen monitoring points from silkworm industry points includes: obtaining the output value and the incidence of silkworm disease of multiple silkworm industry points in the previous year, calculating the ratio of the incidence to the output value, and obtaining the first loss rate of the silkworm industry point due to silkworm disease. The output value of the silkworm industry point can be used to evaluate its breeding scale, business scale and economic benefits. The higher the output value, the larger the breeding scale and the more frequent the business activities. If a silkworm disease outbreak occurs, the greater the loss caused. Therefore, it is representative and effective to use it as a pathogen monitoring point, and more samples to be tested can also be collected, so that the accuracy of the pathogen prediction model is improved. The first loss rate is sorted, and multiple silkworm industry points with a high ranking are obtained as the pathogen monitoring points. Using the ratio of the incidence of silkworm disease and the output value as the first loss rate can intuitively reflect the degree of influence of pathogens on silkworm industry points, so as to quickly screen out representative silkworm industry points, which helps to concentrate resources and attention and improve monitoring efficiency.

[0041] Since pathogens can be transmitted through contact, the personnel exchanges at the silkworm industry sites during the breeding process are prone to cross-infection of pathogens, increasing the risk of being affected by pathogens. Therefore, as the preferred technical solution of this application, after determining the pathogen monitoring site, it also includes: obtaining the number of roads and traffic flow between the pathogen monitoring site and other silkworm industry sites. Preferably, the traffic flow is related to silkworm industry activities, such as silkworm seed transportation, mulberry leaf transportation, silkworm industry exchanges, silkworm cocoon trading and other activities. The higher the frequency of these activities, the more likely they are to cause the spread and infection of pathogens. Based on the number of roads and the traffic flow, the infection weight of the pathogen monitoring site to other silkworm industry sites is determined, for example, using subjective weight method, entropy weight method and nonlinear regression method. In this embodiment, preferably, the number of roads and traffic flow per day are used as independent variables, and the pathogen concentration is used as the dependent variable to construct a nonlinear regression function; the coefficient of the number of roads and the coefficient of traffic flow are calculated through historical data, and the product of the two coefficients is used as the infection weight. This method can not only analyze the degree of influence of pathogen infection caused by traffic between silkworm industry sites, but also analyze the influence coefficient of the number of roads and traffic flow on pathogens, which can provide a basis for the selection of silkworm industry sites. The silkworm industry sites with infection weights greater than the preset infection threshold are also added to the pathogen monitoring sites.

[0042] Example 3: See Figure 3-Figure 4 As shown, based on the first embodiment, this embodiment further adds the technical feature of joint prevention and control of pathogens at silkworm industry sites.

[0043] As a preferred technical solution of the present application, weather data of multiple silkworm industry sites are obtained, including temperature, humidity, precipitation, wind speed, wind force, etc. The silkworm industry sites are divided into multiple sets based on the weather data. For example, the weather data is divided into multiple temperature ranges, and the silkworm industry site sets with similar temperatures are determined based on the number of days corresponding to the temperature ranges; or based on the standard deviation of the weather data, the silkworm industry sites with similar weather data fluctuations are divided into the same set. The silkworm industry sites in the same set are linked for early warning and prevention and control. During the silkworm breeding cycle, weather is one of the important factors affecting the increase in pathogen value. Among the silkworm industry sites with similar weather, the process of pathogen value increase being affected by weather is also similar. Based on the weather data, these similar silkworm industry sites are determined, and unified prevention and control measures can be taken for these sites. There is no need to perform pathogen detection on each site, which helps to improve the efficiency of pathogen monitoring and help other undetected sites to respond quickly to prevention and control.

[0044] Further, see Figure 3 As shown, the method of dividing the silkworm industry sites into multiple sets based on the weather data includes: obtaining the historical temperature and humidity of multiple silkworm industry sites, and performing cluster analysis on the historical temperature and humidity of each silkworm industry site by methods such as K-means, hierarchical clustering, DBSCAN, etc., to obtain multiple temperature and humidity data sets. The number of data sets can be calculated by the elbow method or determined by silkworm experts based on experience. Calculate the statistical values ​​of the temperature and humidity data sets, such as the mean value, mode, etc., directly combine the statistical values, and encode them into a first clustering vector. Based on the first clustering vector, calculate the similarity between the silkworm industry sites, record it as the first similarity, and divide the silkworm industry sites whose first similarity meets the preset error range into the same set.

[0045] Silkworm industry sites with similar temperature and humidity may have the same pathogen development trend. By clustering the historical temperature and humidity data, silkworm industry sites with similar climate characteristics can be identified, making early warning and prevention and control measures more targeted and improving prevention and control efficiency. Cluster analysis reduces the error of subjective classification and makes decision-making more scientific and objective.

[0046] Further, see Figure 4 As shown, the method of dividing the silkworm industry sites into multiple sets based on the weather data includes: The historical wind direction and wind force of the plurality of silkworm industry sites are obtained. Since the wind direction and wind force are qualitative data, in this embodiment, the number of days with wind direction is counted and the days are combined into a wind direction vector; the number of days with wind force is counted and the days are combined into a wind force vector. The wind direction vector and the wind vector are directly combined and encoded into a second clustering vector. Based on the second clustering vector, the similarity between the silkworm industry sites is calculated, recorded as the second similarity, and the silkworm industry sites whose second similarity meets the preset error range are divided into the same set.

[0047] Silkworm industry locations with similar wind direction and wind force show the same pathogen spread trend with the wind direction. By analyzing historical wind direction and wind force, silkworm industry locations affected by pathogen spread can be identified, which is conducive to analyzing the regional patterns of pathogen spread and improving prevention and control efficiency.

[0048] Embodiment 4: Based on the above embodiments, this embodiment further constructs the pathogen prediction model based on the long short-term memory neural network LSTM, and learns the characteristics of weather data affecting the pathogen proliferation process during the breeding cycle by means of the slow execution characteristics of LSTM. Then, based on the weather forecast data for the next few days, the proliferation results of pathogens in the next few days are predicted. The core components of LSTM include cell units, input gates, forget gates, and output gates.

[0049] The first time series of pathogen concentration and the second time series of weather data are obtained, and the pathogen prediction model is input through the input gate in chronological order to extract the contextual features of the weather data affecting the proliferation of pathogens. For example, there is an optimum growth temperature for the rate of pathogen proliferation. When the temperature is lower than the optimum growth temperature, the rate increases with the increase of temperature; when the temperature is higher than or equal to the optimum growth temperature, the rate decreases with the increase of temperature. Humidity not only affects the water required for pathogen proliferation, but also affects the temperature. Wind force and wind direction affect the regional infection process of pathogens, and also affect temperature and humidity. The cell unit is used to store and learn the contextual features, so that the pathogen prediction model can predict the pathogen proliferation process based on different weather data. The pathogen prediction model deletes the interfering features through the forgetting gate. For example, historical temperature and humidity outside the silkworm breeding cycle; abnormal temperature, etc., reduce the interference in the process of learning contextual features of the pathogen prediction model and improve the prediction accuracy. The pathogen prediction model outputs the pathogen concentration corresponding to the weather forecast data through the output gate, and outputs the contextual features retained in the current cell unit to the next cell unit.

[0050] Weather has a continuous impact on the proliferation of pathogens, and the pathogen proliferation process itself also has a quantitative relationship, which makes the time series of pathogen concentration have a contextual relationship. In order to capture this contextual relationship, it is preferred to build a pathogen prediction model based on the LSTM long short-term memory neural network, and retain the contextual features of the weather data affecting the pathogen proliferation process through cell units, so that the prediction results also have a continuous relationship, which is conducive to accurately reflecting the development trend of pathogen concentration.

[0051] Embodiment 5: This embodiment provides a silkworm disease prevention and control system, device and storage medium based on artificial intelligence to predict pathogen trends.

[0052] See also Figure 5 As shown in the figure, the silkworm disease prevention and control system based on artificial intelligence to predict pathogen trends includes: A sampling module is used to select pathogen monitoring points from silkworm industry points, collect samples to be detected from the pathogen monitoring points based on preset sampling rules, and detect the samples to be detected; the sampling rules include sampling timing rules and sampling fixed point rules; An acquisition module is used to store the obtained detection results and the weather data corresponding to the pathogen detection in the pathogen database, and to feed back the detection results to the pathogen monitoring point; set the pathogen type to be predicted and the starting prediction time, and obtain the weather forecast data for the next several days starting from the starting prediction time; The prediction module inputs the weather forecast data into the pathogen prediction model to obtain the pathogen concentration of the pathogen type in the future multiple days, which is recorded as a first prediction content set; the first prediction content set is classified, and early warning and prevention and control measures are determined according to the classification results.

[0053] A silkworm disease prevention and control device based on artificial intelligence prediction of pathogen trends, the device comprising: network hardware for collecting various data of silkworm disease pathogens detected in different places; at least one processor, capable of processing various data related to silkworm disease pathogens collected by network hardware; at least one memory, for storing various data related to silkworm disease pathogens collected by sensors and at least one program; when the at least one program is executed by the at least one processor, so that the at least one processor implements the silkworm disease prevention and control method, the processor and the memory can be connected to store at least one program through a bus or other means; 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 through 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, so that the at least one processor implements the silkworm disease prevention and control method.

[0054] A computer storage medium stores a program executable by a processor, wherein the program executable by the processor is used to implement the silkworm disease prevention and control method when executed by the processor.

[0055] 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.

[0056] 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 method for preventing and controlling silkworm diseases based on artificial intelligence prediction of pathogen trends, characterized in that: include: Select pathogen monitoring points from silkworm industry points, collect samples to be tested from the pathogen monitoring points based on preset sampling rules, and test the samples to be tested; the sampling rules include sampling timing rules and sampling fixed point rules; The obtained detection results and the weather data corresponding to the pathogen detection are stored in the pathogen database, and the detection results are fed back to the pathogen monitoring point; the type of the pathogen to be predicted and the starting prediction time are set, and the weather forecast data for the next several days is obtained starting from the starting prediction time; The weather forecast data is input into the pathogen prediction model to obtain the pathogen concentration of the pathogen type in the future multiple days, which is recorded as a first predicted content set; the first predicted content set is classified, and early warning and prevention and control measures are determined according to the classification results.

2. The method for preventing and controlling silkworm diseases based on artificial intelligence prediction of pathogen trends according to claim 1, characterized in that: The step of grading the first predicted content set comprises: The pathogen concentration is graded, and the pathogen concentration corresponding to the first predicted content set is converted into a first graded set; the number of each grade in the first graded set is counted, and the severity of the pathogen in the first graded set is determined according to the number of the grades.

3. The method for preventing and controlling silkworm diseases based on artificial intelligence prediction of pathogen trends according to claim 2, characterized in that: Before classifying the first classified set, the method further includes smoothing the first classified set, the method comprising: Preset a first threshold and a second threshold; obtain a first subset whose number of grades is greater than the first threshold, and the grades in the first subset are the same and continuous in time; obtain a second subset adjacent to the first subset, if the number of continuous grades in the second subset is less than or equal to the second threshold, regard the second subset as a discrete data set, and modify the second subset so that its grade is the same as that of the first subset.

4. The method for preventing and controlling silkworm diseases based on artificial intelligence prediction of pathogen trends according to claim 2, characterized in that: The step of grading the pathogen concentration comprises: A first concentration threshold and a second concentration threshold are preset; the pathogen concentration is classified as low concentration if it is less than or equal to the first concentration threshold, as medium concentration if it is greater than the first concentration threshold and less than or equal to the second concentration threshold, and as high concentration if it is greater than the second concentration threshold.

5. The method for preventing and controlling silkworm diseases based on artificial intelligence prediction of pathogen trends according to claim 1, characterized in that: The method for selecting pathogen monitoring points from silkworm industry points includes: Obtain the output value and the incidence rate of silkworm diseases of multiple silkworm industry sites in the previous year, calculate the ratio of the incidence rate to the output value, and obtain the first loss rate of the silkworm industry sites caused by silkworm diseases; sort the first loss rates, and obtain multiple silkworm industry sites with top rankings as the pathogen monitoring sites.

6. The method for preventing and controlling silkworm diseases based on artificial intelligence prediction of pathogen trends according to claim 5, characterized in that: After determining the pathogen monitoring point, the method further includes: Obtain the number of roads and traffic flow between the pathogen monitoring point and other silkworm industry points; determine the infection weight of the pathogen monitoring point to other silkworm industry points based on the number of roads and the traffic flow; and add the silkworm industry points whose infection weight is greater than a preset infection threshold to the pathogen monitoring point.

7. The method for preventing and controlling silkworm diseases based on artificial intelligence prediction of pathogen trends according to claim 1, characterized in that: include: Acquire weather data of a plurality of silkworm industry sites, and divide the silkworm industry sites into a plurality of sets based on the weather data; Carry out joint early warning and prevention and control for the silkworm industry locations in the same collection.

8. The method for preventing and controlling silkworm diseases based on artificial intelligence prediction of pathogen trends according to claim 7, characterized in that: The method of dividing the silkworm industry sites into a plurality of sets based on the weather data comprises: Acquire the historical temperature and humidity of the plurality of silkworm industry sites, perform cluster analysis on the historical temperature and humidity of each silkworm industry site to obtain a plurality of temperature and humidity data sets, and encode the temperature and humidity data sets into a first clustering vector; The similarity between the silkworm industry sites is calculated based on the first clustering vector, recorded as the first similarity, and the silkworm industry sites whose first similarity satisfies a preset error range are divided into the same set.

9. The method for preventing and controlling silkworm diseases based on artificial intelligence prediction of pathogen trends according to claim 7, characterized in that: The method of dividing the silkworm industry sites into a plurality of sets based on the weather data comprises: Obtaining historical wind directions and wind forces of a plurality of said silkworm industry points, counting the number of days with wind direction, and combining the number of days into a wind direction vector; counting the number of days with wind force, and combining the number of days into a wind force vector; directly combining the wind direction vector and the wind vector, and encoding them into a second clustering vector; The similarity between the silkworm industry sites is calculated based on the second clustering vector, recorded as the second similarity, and the silkworm industry sites whose second similarity meets the preset error range are divided into the same set.

10. A silkworm disease prevention and control system based on artificial intelligence to predict pathogen trends, characterized in that: include: A sampling module is used to select pathogen monitoring points from silkworm industry points, collect samples to be detected from the pathogen monitoring points based on preset sampling rules, and detect the samples to be detected; the sampling rules include sampling timing rules and sampling fixed point rules; An acquisition module is used to store the obtained detection results and the weather data corresponding to the pathogen detection in the pathogen database, and to feed back the detection results to the pathogen monitoring point; set the pathogen type to be predicted and the starting prediction time, and obtain the weather forecast data for the next several days starting from the starting prediction time; The prediction module inputs the weather forecast data into the pathogen prediction model to obtain the pathogen concentration of the pathogen type in the future multiple days, which is recorded as a first prediction content set; the first prediction content set is classified, and early warning and prevention and control measures are determined according to the classification results.

Images