Smart plant cultivation service system through pests and weed prediction

Abstract

This invention relates to a smart plant cultivation service system through eco-friendly pest and weed occurrence forecasting, which analyzes information on the facility's plant cultivation environment and the condition of cultivated plants based on artificial intelligence (AI) to forecast pest and weed occurrences, predict the spread rate of pests and weeds and plant nutritional status, and propose control and treatment measures based on these results to reduce the use of pesticides, fertilizers, and herbicides. The system includes a sensor that acquires information on the condition of cultivated plants in the facility and surrounding environmental information, and a smart plant cultivation service server that forecasts pest and weed occurrences and predicts the spread rate of pests and weeds and plant nutritional status through AI-based analysis based on information collected through the sensor and various other information on pests and weeds obtained through servers of public agricultural institutions, and proposes control and treatment measures based on these results. The smart plant cultivation service server comprises a data processing unit that preprocesses collected cultivated plant image data on a frame-by-frame basis, a classifying unit that generates a dataset by classifying the preprocessed data, processes the generated dataset with a prediction model to generate prediction results, and uses the generated prediction results to control and A smart plant cultivation service system is implemented through the prediction of pest and weed occurrences, comprising a data analysis unit that generates treatment plans and an analysis information provision unit that provides the generated control and treatment plans as analysis information.

Description

Technology Field

[0001] The present invention relates to a smart plant cultivation service system through the prediction of pest and weed occurrences. In particular, it relates to an eco-friendly smart plant cultivation service system through the prediction of pest and weed occurrences that analyzes information on the plant cultivation environment and the condition of cultivated plants in a facility based on artificial intelligence (AI) to predict the occurrence of pests and weeds and the spread rate of pests and weeds, and suggests control and treatment methods based on this to reduce the use of pesticides, fertilizers, herbicides, etc. Background Technology

[0002] In modern society, population growth and economic development have necessitated increased production of food resources, and significant efforts have been made to introduce new varieties or intensive cultivation techniques to achieve this. Meanwhile, with the introduction of various varieties and the use of new chemicals for controlling pests, diseases, and weeds, the patterns of disease, pest, and weed occurrence have become more complex, leading to environmental pollution problems caused by the use of pesticides and fertilizers, as well as issues regarding residual toxicity of pesticides.

[0003] Therefore, since crop diseases, pests, and weeds have a significant impact on agricultural production, it is necessary to analyze their causes and prevent them in advance; furthermore, it is important to create optimal conditions for plant growth in environments such as global warming and abnormal temperatures.

[0004] Currently, it is difficult for farmers in domestic greenhouses to diagnose subtle early symptoms of pests and diseases, and there is a problem in that the labor cost burden on farms is too high to deploy manpower daily to check for these early symptoms.

[0005] In this situation where accurate diagnosis is difficult, excessive pest control based on farmers' experience is being carried out, leading to concerns regarding pesticide residues.

[0006] Meanwhile, monitoring refers to the activity of observing the occurrence of pests, diseases, and weeds in advance, predicting potential problems beforehand, and resolving them.

[0007] Currently, the application of pesticides and fertilizers in rural areas is carried out by relying on the direct experience of farmers or workers to apply them in advance or after pests and weeds have appeared, rather than through concepts such as the aforementioned monitoring.

[0008] This method has the problem that it is a reactive process carried out after pests and weeds have occurred in crops, and also has the problem of polluting the environment by overusing pesticides and fertilizers through preemptive spraying in environments where they have not yet occurred. Prior art literature

[0009] Republic of Korea Registered Patent 10-2655126 (Registered Apr. 02, 2024) (Smart Preventive Agriculture Service System through AI Technology-based Disease Surveillance) Republic of Korea Published Patent 10-2025-0016746 (Published Feb. 04, 2025) (AI-based Cloud Service System for Horticultural Disease Diagnosis and Prescription Support) The problem to be solved

[0010] Therefore, the present invention is proposed to solve the various problems occurring in general greenhouse plant cultivation devices as described above. Its purpose is to provide a smart plant cultivation service system through environmentally friendly pest and weed occurrence forecasting, which analyzes information on the facility's plant cultivation environment and the condition of cultivated plants based on artificial intelligence (AI) to forecast the occurrence of pests and weeds and predict their spread speed, and based on this, suggests control and treatment methods to reduce the use of pesticides, fertilizers, and herbicides. means of solving the problem

[0011] In order to achieve the above-mentioned purpose, the first embodiment of the "smart plant cultivation service system through pest and weed occurrence prediction" according to the present invention is,

[0012] A sensor for acquiring status information of cultivated plants in the facility and surrounding environment information; and

[0013] It includes a smart plant cultivation service server that predicts the occurrence of pests and weeds and the spread speed of pests and weeds, predicts plant nutritional status, and suggests control and treatment methods based on this, through artificial intelligence (AI)-based analysis based on information collected through the above-mentioned sensors and plant condition information and pest and disease information obtained through servers of public agricultural institutions.

[0014] The above smart plant cultivation service server is,

[0015] A data collection unit that collects sensor-collected information and agriculture-related information through servers of public institutions related to agriculture;

[0016] A data processing unit that preprocesses image data collected by the above data collection unit in frame units;

[0017] A data analysis unit that classifies data preprocessed by the above data processing unit to generate a dataset, processes the generated dataset with a prediction model to generate prediction results, and generates control and treatment plans based on the generated prediction results;

[0018] It is characterized by including an analysis information providing unit that provides pest control and treatment plans generated by the above-mentioned data analysis unit as analysis information.

[0019] Preferably, the data analysis unit is,

[0020] A dataset generation unit that constructs a dataset by labeling cultivated plant status information collected from a sensor;

[0021] A model generation unit that generates a pest prediction model for pest monitoring, a pest spread speed prediction model for predicting the spread speed of pests, a weed state prediction model for predicting weed state, and a plant nutrient state prediction model for predicting plant nutrient state;

[0022] A model training unit that trains the models generated by the above model generation unit, respectively, based on deep learning;

[0023] It is characterized by including a model optimization unit that optimizes a pest and disease prediction model, a pest and disease spread speed prediction model, a weed condition prediction model, and a plant nutrient condition prediction model based on the learning results of the above-mentioned model learning unit.

[0024] Preferably, the data analysis unit is,

[0025] A prediction model application unit that applies the above-mentioned constructed dataset to prediction models generated for the forecasting of pests and diseases of cultivated plants, the prediction of their spread speed, the prediction of weed status, and the prediction of plant nutritional status, respectively;

[0026] It is characterized by including a prediction result generation unit that generates a prediction result based on the output result of the prediction model application unit and weather information.

[0027] Preferably, the prediction model application unit is,

[0028] A pest and disease prediction model that performs pest and disease forecasting based on the condition information of cultivated plants;

[0029] A pest spread rate prediction model that predicts the spread rate of pests based on the condition information of cultivated plants;

[0030] A weed state prediction model that predicts weed state based on the state information of the cultivated plants mentioned above;

[0031] It is characterized by including a plant nutritional status prediction model that predicts the plant nutritional status based on the state information of the cultivated plant.

[0032] Preferably, the data analysis unit is,

[0033] A pest and disease control and treatment plan generation unit that generates control and treatment plans by searching a database based on the pest and disease prediction results, pest and disease spread speed prediction results, or weed condition prediction results generated by the above prediction result generation unit;

[0034] It is characterized by including an analysis information providing unit that outputs the pest control and treatment plan generated by the above pest control and treatment plan generating unit as analysis information.

[0035] Preferably, the above-mentioned pest control and treatment method generating unit is,

[0036] A pest determination unit that determines pests based on pest prediction results or pest spread speed prediction results;

[0037] It is characterized by including a pest and disease control and treatment plan determination unit that derives pest and disease control and treatment plans pre-registered in a database based on pest and disease control information determined by the above-mentioned pest and disease control determination unit.

[0038] Preferably, the above-mentioned pest control and treatment plan determining unit is,

[0039] It is characterized by determining whether to use pesticides and fertilizers and the amount to use based on pest and disease prediction results and pest and disease spread speed prediction results.

[0040] Preferably, the above-mentioned pest control and treatment method generating unit is,

[0041] A weed state determination unit that determines the weed state based on the weed state prediction result;

[0042] It is characterized by including a weed treatment plan determination unit that derives a weed treatment plan previously registered in a database based on the weed status determined by the weed status determination unit above.

[0043] Preferably, the weed treatment method determining unit is,

[0044] It is characterized by converting the weed status determined by the weed status determination unit into weights, and applying the converted weights to an impact analysis model on pests and plant cultivation to determine whether to use herbicides and the amount of herbicides to use.

[0045] Preferably, the above-mentioned pest control and treatment method generating unit is,

[0046] A plant nutritional status verification unit that verifies the plant nutritional status based on the above plant nutritional status prediction results;

[0047] It is characterized by including a nutrient status treatment plan determination unit that determines the supply amount of nutrients, water, and light for plant growth as a treatment plan based on the plant nutrient status confirmed by the above-mentioned plant nutrient status confirmation unit.

[0048] In order to achieve the above-mentioned purpose, a second embodiment of the "smart plant cultivation service system through pest and weed occurrence prediction" according to the present invention is,

[0049] It includes a smart terminal that predicts the occurrence of pests and weeds, the spread rate of pests and weeds, and plant nutritional status through artificial intelligence (AI)-based analysis based on information collected through sensors acquiring status information of cultivated plants in the facility and surrounding environment information, as well as plant status information and pest and disease information acquired through servers of public agricultural institutions, and suggests control and treatment methods based on this.

[0050] The above smart terminal is,

[0051] A data collection unit that collects sensor-collected information and agriculture-related information through servers of public institutions related to agriculture;

[0052] A data processing unit that preprocesses image data collected by the above data collection unit in frame units;

[0053] A data analysis unit that classifies data preprocessed by the above data processing unit to generate a dataset, processes the generated dataset with a prediction model to generate prediction results, and generates control and treatment plans based on the generated prediction results;

[0054] It is characterized by including an analysis information providing unit that provides pest control and treatment plans generated by the above-mentioned data analysis unit as analysis information. Effects of the invention

[0055] According to the present invention, by analyzing the plant cultivation environment and condition information of cultivated plants in a facility (particularly a greenhouse) based on artificial intelligence (AI), it is possible to predict the occurrence of pests and weeds and the spread rate of pests and weeds, and to predict the nutritional status of plants. Based on this, control and treatment methods are proposed to minimize the use of pesticides, fertilizers, herbicides, etc., thereby providing an environmentally friendly control and treatment method. Brief explanation of the drawing

[0056] FIG. 1 is a schematic diagram of a first embodiment of a smart plant cultivation service system through pest and weed occurrence prediction according to the present invention, and FIG. 2 is a configuration diagram of an exemplary embodiment of the data analysis unit of FIG. 1, and Figure 3 is an example of a prediction model used in the prediction model application section of Figure 2, and FIG. 4 is a configuration diagram of an exemplary embodiment of the pest control and treatment plan generation unit of FIG. 2, and FIG. 5 is a schematic diagram of a second embodiment of a smart plant cultivation service system through pest and weed occurrence prediction according to the present invention. Specific details for implementing the invention

[0057] A smart plant cultivation service system for predicting pest and weed occurrence according to a preferred embodiment of the present invention will be described in detail below with reference to the attached drawings.

[0058] The terms or words used in the present invention described below should not be interpreted as being limited to their ordinary or dictionary meanings, but should be interpreted in a meaning and concept consistent with the technical spirit of the present invention, based on the principle that the inventor can appropriately define the concept of the terms to best describe his invention.

[0059] Therefore, the embodiments described in this specification and the configurations illustrated in the drawings are merely preferred embodiments of the present invention and do not represent all technical concepts of the present invention; thus, it should be understood that various equivalents and modifications that can replace them may exist at the time of filing this application.

[0060] <Example 1>

[0061] FIG. 1 is a schematic diagram of a smart plant cultivation service system through pest and weed occurrence prediction according to a preferred first embodiment of the present invention, and may include a sensor (100), an agricultural public institution server (200), and a smart plant cultivation service server (300).

[0062] The sensor (100) plays the role of acquiring status information of cultivated plants and surrounding environment information of a facility (which may be a greenhouse or a smart farm as a plant cultivation facility) and uploading it to a smart plant cultivation service server (300) in real time. This sensor (100) can acquire status information of cultivated plants by photographing the inside of the facility using unmanned mobile equipment such as a drone equipped with a camera or a camera (CCTV) installed at a specific location. In addition, it can acquire temperature, humidity, illuminance, etc. inside the facility using a temperature sensor, a humidity sensor, an illuminance sensor, etc.

[0063] The agricultural public institution (200) is a server that provides information on pests and other various things, and can receive information on agricultural pests, weeds, plant nutritional status, and surrounding weather information, temperature, humidity, etc.

[0064] The smart plant cultivation service server (300) predicts the occurrence of pests and weeds and the spread speed and nutritional status of plants through artificial intelligence (AI)-based analysis based on information collected through the sensor (100) and information obtained through the server of an agricultural public institution (200), and provides control and treatment methods based on this.

[0065] The smart plant cultivation service server (300) of the present invention aims to implement an eco-friendly plant cultivation service by reducing the use of pesticides, fertilizers, and herbicides by comprehensively analyzing the growth status of plants, the status of weeds, and plant growth factors caused by the surrounding environment, rather than presenting reactive pest control and treatment methods, and by predicting pest and disease monitoring, the spread speed of pests and diseases, and the condition of weeds, and providing information on the use of pesticides, fertilizers, and herbicides in advance. In addition, it aims to optimize plant growth by checking the nutritional status of plants and, based on this, providing information on nutrients, water, light intensity, etc. for plant growth.

[0066] In particular, by analyzing video images within the facility to assess the condition of weeds and analyzing the impact of the weed conditions on plant growth, weeding can be performed in advance to prevent factors inhibiting plant growth and reduce the use of herbicides, thereby promoting an environmentally friendly plant cultivation service.

[0067] This smart plant cultivation service server (300) may include a data collection unit (310) that collects sensor collection information and agricultural-related information, a data processing unit (320) that preprocesses image data collected by the data collection unit (310) in frame units, a data analysis unit (340) that classifies the data preprocessed by the data processing unit (320) to create a dataset, processes the created dataset with a deep learning prediction model to create a prediction result, and creates a pest control and treatment plan based on the created prediction result, an analysis information provision unit (350) that provides the pest control and treatment plan created by the data analysis unit (340) as analysis information, and a database (330) in which the pest control and treatment plan is registered.

[0068] As illustrated in FIG. 2, the data analysis unit (340) may include a collected data storage unit (341) that collects and stores data from the sensor (100) and an agricultural public institution (200), a dataset generation unit (342) that constructs a dataset by labeling the cultivated plant status information among the collected data, a model generation unit (344) that generates a pest prediction model for pest monitoring, a pest spread speed prediction model for predicting the pest spread speed, a weed state prediction model for predicting the weed state, and a plant nutrient state prediction model for predicting the plant nutrient state, a model learning unit (345) that learns the models generated by the model generation unit (344) based on deep learning, and a model optimization unit (346) that optimizes the pest prediction model, the pest spread speed prediction model, the weed state prediction model, and the plant nutrient state prediction model based on the learning results of the model learning unit (345).

[0069] Additionally, the data analysis unit (340) may include a prediction model application unit (343) that applies the constructed dataset to a prediction model created for the prediction of pests and diseases of cultivated plants and the prediction of their spread speed, as well as for the prediction of weeds and the prediction of their spread speed and the prediction of their nutritional status, and a prediction result generation unit (347) that generates a prediction result based on the output result of the prediction model application unit (343) and agricultural information.

[0070] Here, the prediction model application unit (343) may include a pest prediction model (411) that performs pest and disease prediction based on the condition information of the cultivated plant as shown in FIG. 3, a pest and disease spread speed prediction model (412) that predicts the spread speed of pests and weeds based on the condition information of the cultivated plant, a weed state prediction model (413) that predicts the weed state and the spread speed of the weed state based on the condition information of the cultivated plant, and a plant nutritional state prediction model (414) that predicts the nutritional state of the plant based on the condition information of the cultivated plant.

[0071] Additionally, the data analysis unit (340) may include a pest and disease prediction result, a pest and disease spread speed prediction result, a weed condition prediction and spread speed result, or a plant nutritional status prediction result generated by the prediction result generation unit (347), a pest and disease control and treatment method generation unit (348) that searches the database (330) to generate a pest and disease control and treatment method, and a collected data storage unit (349) that stores the pest and disease control and treatment method generated by the pest and disease control and treatment method generation unit (348) as analysis information.

[0072] As shown in FIG. 4, the above pest control and treatment plan generation unit (348) may include a pest determination unit (421) that determines pests based on pest prediction results or pest spread speed prediction results, and a pest control and treatment plan determination unit (422) that derives pest control and treatment plans previously registered in a database (330) based on pest determination information determined by the pest determination unit (421).

[0073] Here, the pest control and treatment method decision unit (422) can determine whether to use pesticides and fertilizers and the amount of pesticides and fertilizers to use based on the pest prediction results and the pest spread speed prediction results.

[0074] Additionally, the above-mentioned pest control and treatment plan generation unit (348) may include a weed state determination unit (423) that determines the weed state based on the weed state prediction result, and a weed treatment plan determination unit (424) that derives a weed treatment plan previously registered in the database (330) based on the weed state determined by the weed state determination unit (423).

[0075] The above weed treatment method determination unit (424) can convert the weed condition determined by the weed condition determination unit (423) into a weight, and apply the converted weight to an impact analysis model on pests and plant cultivation to determine whether to use herbicides and the amount of herbicides to use.

[0076] Additionally, the above-mentioned pest control and treatment plan generation unit (348) may include a plant nutritional status verification unit (425) that verifies the plant nutritional status based on the plant nutritional status prediction result, and a nutritional status treatment plan determination unit (426) that determines the supply amount of nutrients, water, and light for plant growth as a treatment plan solution based on the plant nutritional status verified by the plant nutritional status verification unit (425).

[0077] The operation of the smart plant cultivation service system through pest and weed occurrence prediction according to the preferred first embodiment of the present invention configured as described above is specifically explained as follows.

[0078] First, the smart plant cultivation service server (300) promotes the growth of plants grown in the facility, prevents pests and diseases in advance, predicts the spread speed when pests and diseases occur, analyzes the condition of weeds to determine the impact on plant growth, and suggests treatment methods such as whether or not to use herbicides, and checks the nutritional status of the plants to suggest the supply amount of nutrients, water, light, etc. for plant growth.

[0079] The model generation unit (344) generates a pest prediction model for pest monitoring, a pest spread speed prediction model for predicting the spread speed of pests, a weed state prediction model for predicting weed state, and a plant nutrient state prediction model for predicting plant nutrient state based on deep learning.

[0080] To achieve this, a method called 'Machine Learning' is utilized. Machine learning is a technology that inputs a large amount of collected data into a model and has it classify similar items together. For example, if similar photos of weeds are input, the model is instructed to classify them as weeds.

[0081] Many machine learning algorithms have already emerged regarding how to classify such image data, and deep learning is a machine learning method proposed to overcome the limitations of artificial neural networks.

[0082] Deep learning mimics the human neural network to create a deep neural network, which is a network built by stacking layers between inputs and outputs to learn and solve combined weight problems. The deep learning used here demonstrates strength in high-level abstraction of objects, which was the biggest weakness of existing machine learning techniques, through a combination of various non-linear transformation methods.

[0083] In an embodiment of the present invention, AI-based prediction of pest and weed occurrence using deep learning, prediction of the spread speed of pests and weeds and the nutritional status of plants are performed, and control and treatment methods are presented accordingly.

[0084] Here, by generating each model based on growth-oriented deep learning, model optimization can be achieved by training based on pre-collected big data, or the models can be continuously upgraded by extracting training data based on real-time cultivated plant status and weather information obtained from the field and using it to train each model independently. Model upgrades enable more accurate performance of pest and disease monitoring, prediction of pest and disease spread speed, weed status, and plant nutrient status prediction.

[0085] The four models generated in the model generation unit (344) each learn based on the collected information obtained in advance through the model learning unit (345).

[0086] For model training, collected plant cultivation videos from the facility are processed frame by frame. Keypoint scores are extracted from the frame-by-frame processed video data using an autoencoder, and four models are trained using the extracted keypoints. Here, keypoint detection classifies cultivated plants into leaves, stems, flowers, colors, etc., based on the cos-similarity between images within the acquired video, and extracts and matches keypoints through comparison between different images. Keypoint scoring (KPS), which represents the similarity between each image, is calculated, and each image (leaf, stem, flower, etc.) is classified using the calculated keypoint scores. Finally, the classified images are trained.

[0087] Based on the learning results obtained by learning each model in this way, the model optimization unit (346) optimizes each model by gradually increasing the accuracy of the model. In addition, the optimized pest and disease prediction model, pest and disease spread speed prediction model, weed condition prediction model, and plant nutritional condition prediction model are provided to the prediction model application unit (343). As noted above, the four models applied to the prediction model application unit (343) can be continuously upgraded based on the cultivated plant condition information and environmental information, such as weather information, obtained in real time from the field.

[0088] With the models for pest and disease monitoring, pest and disease spread speed prediction, weed condition monitoring and weed spread speed prediction, and plant nutritional status prediction established as described above, when the actual plant cultivation system is operated, the sensor (100) acquires plant status information (plant image) and surrounding environment information of the facility (which may be a greenhouse or a smart farm as a plant cultivation facility) and uploads them to the smart plant cultivation service server (300) in real time.

[0089] In addition, the smart plant cultivation service server (300) acquires various agricultural information and weather information through the server of an agricultural public institution (200).

[0090] Based on information collected from facilities in the field and various information obtained through servers of public institutions related to agriculture, we predict the occurrence of pests and weeds, the spread speed of pests and weeds, and the nutritional status of plants through AI-based deep learning analysis, and based on this, we propose control and treatment methods.

[0091] For example, the smart plant cultivation service server (300) comprehensively analyzes the growth status of plants, the status of weeds, and plant growth factors due to the surrounding environment rather than presenting post-event pest control and treatment methods, predicts pest and disease detection and the spread speed of pests and diseases, weed detection and the spread speed of weeds, and predicts the nutritional status of plants, thereby providing information on the use of pesticides, fertilizers, and herbicides in advance to reduce the use of pesticides, fertilizers, and herbicides, thereby implementing an eco-friendly plant cultivation service.

[0092] In particular, by analyzing video images within the facility to assess the condition of weeds and their impact on plant growth, it enables preemptive weeding to prevent factors that inhibit plant growth and reduce herbicide use, thereby promoting an environmentally friendly plant cultivation service.

[0093] In addition, by predicting plant nutritional status, it suggests the supply amounts of nutrients, water, and light necessary for plant growth, thereby promoting optimal growth.

[0094] In other words, the data collection unit (310) of the smart plant cultivation service server (300) collects sensor collection information and information from public institutions related to agriculture, and the data processing unit (320) preprocesses the collected image data frame by frame and stores it in the database (330).

[0095] The data analysis unit (340) classifies the data preprocessed by the data processing unit (320) to generate a dataset, processes the generated dataset with four deep learning prediction models to generate prediction results, and generates pest control and treatment methods based on the generated prediction results.

[0096] That is, the data analysis unit (340) stores the collected data from the sensor (100) and the agricultural-related public institution server (200), which has been preprocessed through the data processing unit (320), in the database (330).

[0097] Next, the dataset generation unit (342) constructs a dataset by labeling the cultivated plant status information among the collected data. Here, labeling refers to the process of classifying image data, for example, into leaves, stems, flowers, weeds, etc., and attaching tags.

[0098] The dataset constructed in this way is applied to each model in the prediction model application unit (343) to predict pests and diseases of cultivated plants and their spread speed, predict weed occurrence and weed spread speed, and predict the nutritional status of plants.

[0099] For example, the pest and disease prediction model (411) of the prediction model application unit (343) generates an output using each image classified from the state information of the cultivated plant as input data. For the leaf image of the classified image, it outputs difference information such as the presence or absence of size and color difference by comparing it with a reference leaf image based on the growth date; for the stem image, it outputs difference information such as the presence or absence of size and color difference by comparing it with a reference stem image based on the growth date; and for the flower image, it outputs difference information such as the presence or absence of size and color difference by comparing it with a reference flower image based on the growth date. This difference information is subsequently quantified to become information that can determine plant growth information and determine the presence or absence of pests and diseases.

[0100] Likewise, the pest spread speed prediction model (411) compares the leaf image, stem image, flower image, etc., of the image classified as the state information of the cultivated plant with a reference image for each image to extract the differences, and predicts the pest spread speed based on the extracted differences. That is, the pest differences extracted through the model are quantified and tracked over a certain period of time or duration, or the pest spread speed is predicted using the current numerical score.

[0101] In addition, by using a weed condition prediction model (413), weed images extracted as condition information of cultivated plants are compared with reference weed images to analyze the presence or absence of weeds, the size of weeds, etc., and by quantifying the results of this analysis, weed occurrence can be predicted and the speed of weed spread can be predicted.

[0102] In addition, the plant nutritional status prediction model (414) compares the status information of the cultivated plant with the standard status information of the cultivated plant based on the current cultivation date to extract the plant growth status and predict the plant nutritional status based on this.

[0103] When pest and disease prediction output, pest and disease spread speed prediction, weed occurrence prediction and weed spread speed prediction, and plant nutritional status prediction are performed through each model via the prediction model application unit (343), the prediction result generation unit (347) generates a prediction result based on the output result of the prediction model application unit (343) and information provided by public institutions related to agricultural pests, and transmits it to the pest control and treatment method generation unit (348).

[0104] The above pest control and treatment plan generation unit (348) generates a pest control and treatment plan by searching the database (330) based on the pest prediction result, pest spread speed prediction result, weed occurrence prediction result, weed condition spread speed prediction result, or plant nutritional status prediction result transmitted from the above prediction result generation unit (347).

[0105] For example, the pest determination unit (421) of the pest control and treatment plan generation unit (348) determines the pest based on the pest prediction results or the pest spread speed prediction results. That is, it determines whether a pest has occurred and what kind of pest has occurred.

[0106] The pest control and treatment plan determination unit (422) derives pest control and treatment plans that are previously registered in the database (330) based on the pest determination information determined by the pest determination unit (421) and stores them in the collected data storage unit (350).

[0107] Here, the pest control and treatment method determination unit (422) predicts the pest and disease prediction results, that is, what pests and diseases are currently occurring or what pests and diseases are expected to occur in the future, and determines whether to use pesticides and fertilizers and the amount of pesticides and fertilizers to use as a pest and disease control and treatment method based on the prediction results of the spread speed of pests and diseases when they occur. Here, if there is a concern about the occurrence of pests and diseases or if the spread speed of pests and diseases that have occurred is predicted, the optimal pest and disease control and treatment method is generated and provided, such as preemptive pest control methods, that is, spraying pesticides or fertilizers, supplying water, and supplying sunlight. This allows for the minimization of the amount of pesticides or fertilizers used compared to post-treatment when pests and diseases occur or become widespread in the future, thereby providing an environmentally friendly plant cultivation method. In particular, by providing information such as sunlight levels, time, temperature, and moisture levels to control or prevent the spread of pests, pest control becomes possible, thereby providing an eco-friendly plant cultivation method that enables pest control without the use of pesticides or fertilizers.

[0108] In addition, the weed condition determination unit (423) determines the weed condition based on the weed occurrence monitoring results and the weed spread speed prediction results. That is, it determines whether weeds have occurred, the size of the weeds that have occurred, and the type of weeds.

[0109] When the condition of the weeds is determined in this way, the weed treatment method determination unit (424) derives a weed treatment method previously registered in the database (330) based on the determined weed condition and stores it in the collected data storage unit (349) as a treatment method.

[0110] Here, the weed treatment method decision unit (424) can convert the weed condition determined by the weed condition determination unit (423) into a weight, and apply the converted weight to a model for analyzing the impact on pests and plant cultivation to determine whether to use herbicides and the amount of herbicides to use.

[0111] In other words, it provides information on the type of herbicide and the amount to be used for weed control in response to the size or type of weeds that have emerged, thereby taking preventive measures to stop weeds from adversely affecting cultivated plants and enabling them to grow optimally.

[0112] In addition, the plant nutritional status verification unit (425) verifies the plant nutritional status based on the predicted results of the plant nutritional status. That is, it verifies whether the nutritional status of the plant is appropriate by comparing standard plant growth status information according to the cultivation period of the plant.

[0113] The information obtained from checking the plant's nutritional status is transmitted to the nutritional status processing method determination unit (426), and the nutritional status processing method determination unit (426) determines a processing method to reduce the supply of nutrients, water, and light if the plant has grown excessively based on the results of checking the nutritional status, or determines a processing method to increase the supply of nutrients, water, and light if it is determined that the plant has grown underdeveloped. Then, to optimize the plant's nutritional status, the determined processing method is stored in the collected data storage unit (349).

[0114] The pest control and treatment plan generated by the pest control and treatment plan generation unit (348) stored in the collected data storage unit (349) is transmitted to the analysis information provision unit (350) as analysis information.

[0115] The above analysis information providing unit (350) provides the pest control and treatment plan generated by the above data analysis unit (340) as analysis information to the user terminal.

[0116] Based on the analysis information of cultivated plants provided in real-time or periodically through the user terminal, the user can suppress the occurrence of pests and diseases in advance by performing pest control treatment, prevent the spread of pests and diseases by spraying pesticides or other treatments when pests and diseases occur, allow the cultivated plants to grow optimally by removing weeds, or appropriately supply and control nutrients, water, and light.

[0117] The above method is a method in which a server links with a sensor (100) and an agricultural public institution server (200) to perform pest and weed occurrence monitoring, check the nutritional status of plants, and then provide control and treatment plans based on the results to a user terminal of a facility farm so that the user can take action. Although this method is sufficient to provide a method for facility farmers to control and treat pests and weeds, it is insufficient for facility farmers to directly perform real-time pest and weed occurrence monitoring at the site, check the nutritional status of plants, and then generate control and treatment plans using this information.

[0118] Accordingly, in another embodiment, the present invention enables a user of a facility farm to directly perform real-time monitoring of pest and weed occurrences at the site, monitor plant nutritional status, and then use this information to generate control and treatment plans.

[0119] <Example 2>

[0120] FIG. 5 is a schematic diagram of a smart plant cultivation service system through pest and weed occurrence prediction according to a preferred second embodiment of the present invention, and may include an agricultural public institution server (200) and a smart terminal (500).

[0121] The agricultural public institution server (200) and the smart terminal (500) can transmit and receive information through the network.

[0122] The agricultural public institution (200) is a server that provides information on pests and other various things, and can receive information on agricultural pests and diseases, as well as surrounding weather information, temperature, humidity information, etc.

[0123] The smart terminal (500) plays a role in predicting the occurrence of pests and weeds and the spread speed of pests and weeds in real time at the site and predicting the nutritional status of plants based on information directly collected using sensors and information obtained through the server of an agricultural public institution (200), and suggesting control and treatment methods based on this.

[0124] These smart terminals (500) can be implemented as mobile devices such as smartphones, laptops with mobile functions, pads, etc.

[0125] The smart terminal (500) of the present invention aims to implement an eco-friendly plant cultivation service by reducing the use of pesticides, fertilizers, and herbicides by comprehensively analyzing the growth status of plants, the condition of weeds, and plant growth factors caused by the surrounding environment, rather than presenting post-hoc control and treatment methods, and by predicting pest and disease monitoring, the spread speed of pests and diseases, and the condition of weeds, and providing information on the use of pesticides, fertilizers, and herbicides in advance.

[0126] In particular, by analyzing video images within the facility to assess weed conditions and plant nutritional status, and by analyzing the impact of the weed conditions on plant growth to enable preemptive weeding, it is possible to prevent factors inhibiting plant growth and reduce herbicide use, thereby promoting an environmentally friendly plant cultivation service.

[0127] In addition, by checking the plant's nutritional status, it is possible to optimize plant growth by providing solutions for the supply of nutrients, water, and light to enable optimal growth.

[0128] This smart terminal (500) may include a sensor (560) for acquiring information on the condition of cultivated plants in a facility and surrounding environment information, a data collection unit (510) for collecting information from the sensor (560) and agricultural-related information, a data processing unit (520) for preprocessing image data collected by the data collection unit (510) in frame units, a data analysis unit (540) for classifying the data preprocessed by the data processing unit (520) to create a dataset, processing the created dataset with a deep learning prediction model to create a prediction result, and creating a pest control and treatment plan based on the created prediction result, an analysis information providing unit (550) for providing the pest control and treatment plan created by the data analysis unit (540) as analysis information, and a database (530) in which the pest control and treatment plan is registered.

[0129] The above data analysis unit (540) is configured with the same configuration as the data analysis unit (340) shown in FIGS. 2, FIGS. 3 and FIGS. 4 and the embodiment configuration, and the operation is also performed in the same way.

[0130] The smart plant cultivation service system for pest and weed occurrence prediction according to the second preferred embodiment of the present invention configured as described above operates and functions in the same way as the smart plant cultivation service system for pest and weed occurrence prediction of FIG. 1 described above, and the difference from FIG. 1 is that the sensor is configured inside the smart terminal (560).

[0131] Here, the sensor (560) may be a device such as a camera, a temperature sensor, a humidity sensor, etc., and when plant information is acquired using a drone inside the facility, the only difference is that the data collection unit (510) acquires the plant cultivation video acquired through the drone through a network with equipment such as the drone.

[0132] Since the operation after acquiring information operates in the same way as in FIG. 1, to avoid redundant explanation, the operation of Embodiment 2 of the present invention will refer to the operation of Embodiment 1 described above.

[0133] In this way, by utilizing smart devices carried by users, it becomes possible to monitor the occurrence of pests and weeds in real time at the facility farm site, predict the spread speed of pests and weeds in real time, and predict plant nutritional status in real time, thereby enabling the optimization of plant cultivation.

[0134] According to the present invention described above, the plant cultivation environment and condition information of cultivated plants in a facility (particularly a greenhouse) are analyzed based on artificial intelligence (AI) to predict the occurrence of pests and weeds and the spread rate of pests and weeds, and to predict the nutritional status of plants. Based on this, control and treatment methods are proposed to minimize the use of pesticides, fertilizers, herbicides, etc., thereby providing environmentally friendly control and treatment methods. Additionally, by providing the supply amounts of nutrients, water, light, etc., optimal plant growth can be achieved.

[0135] Although the invention made by the inventors has been specifically described according to the above embodiments, it is obvious to those skilled in the art that the invention is not limited to the above embodiments and can be modified in various ways without departing from the gist thereof. Explanation of the symbols

[0136] 100, 560: Sensor 200: Agriculture-related public institution server 300: Smart Plant Cultivation Service Server 310, 510: Data Collection Unit 320, 520: Data processing unit 330, 530: Database (DB) 340, 540: Data Analysis Department 341: Collected Data Storage Unit 342: Dataset creation section 343: Prediction Model Application Section 344: Model creation section 345: Model learning section 346: Model Optimization Section 347: Prediction result generation unit 348: Control and Treatment Plan Generation Section 349: Analysis Information Output Section 350: Analysis Information Provision Department

Claims

Claim 1 Sensor for acquiring status information of cultivated plants in the facility and surrounding environment information; The system includes a smart plant cultivation service server that predicts the occurrence of pests and weeds and the spread speed of pests and weeds, and predicts plant nutritional status based on artificial intelligence (AI)-based analysis, using information collected through the sensor and plant condition information and pest information obtained through an agricultural public institution server, and suggests control and treatment methods based thereon. The smart plant cultivation service server includes: a data collection unit that collects sensor-collected information and agricultural information obtained through an agricultural public institution server; a data processing unit that preprocesses the cultivated plant condition information, surrounding environment information, and agricultural information collected by the data collection unit in frame units; a data analysis unit that classifies the data preprocessed by the data processing unit to generate a dataset, processes the generated dataset with a prediction model to generate prediction results, and generates control and treatment methods based on the generated prediction results; and an analysis information provision unit that provides the control and treatment methods generated by the data analysis unit as analysis information. The data analysis unit [prepares] the pest prediction results, pest spread speed prediction results, weed condition prediction results, or plant nutritional status generated by the prediction result generation unit. It includes a pest and treatment plan generation unit that generates pest and treatment plans by searching a database based on prediction results, wherein the pest and treatment plan generation unit comprises: a pest determination unit that determines pests based on pest and disease prediction results or pest and disease spread speed prediction results; a pest and treatment plan determination unit that derives pest and treatment plans previously registered in a database based on pest and disease determination information determined by the pest and disease determination unit; a weed state determination unit that determines weed state based on weed state prediction results; a weed treatment plan determination unit that derives weed treatment plans previously registered in a database based on weed state determined by the weed state determination unit; and a plant nutrient state verification unit that verifies plant nutrient state based on plant nutrient state prediction results.The above includes a nutrient status treatment plan determination unit that determines the supply amount of nutrients, water, and light for plant growth as a treatment plan based on the plant nutrient status confirmed by the above plant nutrient status confirmation unit; the above pest control and treatment plan determination unit determines whether to use pesticides and fertilizers and the amount of pesticides and fertilizers to use based on the pest and disease prediction results and the pest and disease spread speed prediction results; the above weed treatment plan determination unit converts the weed status determined by the weed status determination unit into a weight, and applies the converted weight to a model for analyzing the impact on pests and diseases and plant cultivation to determine whether to use herbicides and the amount of herbicides to use; and the above data analysis unit includes a dataset generation unit that constructs a dataset by labeling cultivated plant status information collected from a sensor; a model generation unit that generates a pest and disease prediction model for pest and disease monitoring, a pest and disease spread speed prediction model for predicting the spread speed of pests and diseases, a weed status prediction model for predicting weed status, and a plant nutrient status prediction model for predicting plant nutrient status; and the models generated by the above model generation unit, respectively A model training unit that trains based on deep learning; a model optimization unit that optimizes a pest prediction model, a pest spread speed prediction model, a weed state prediction model, and a plant nutrient state prediction model based on the training results of the model training unit; a prediction model application unit that applies a constructed dataset to the prediction models generated for pest monitoring, spread speed prediction, weed state prediction, and plant nutrient state prediction of cultivated plants, respectively; and a prediction result generation unit that generates prediction results based on the output results of the prediction model application unit and weather information, wherein the prediction model application unit includes: a pest prediction model that performs pest monitoring based on state information of cultivated plants; a pest spread speed prediction model that predicts the pest spread speed based on state information of cultivated plants; and a weed state prediction model that predicts weed occurrence and weed spread speed based on state information of cultivated plants.It includes a plant nutrient status prediction model that predicts the plant nutrient status based on the state information of the cultivated plants mentioned above, and the weed status prediction model analyzes the presence and size of weeds by comparing weed images extracted from the state information of the cultivated plants with reference weed images, and predicts weed occurrence and the speed of weed spread by quantifying these analysis results. The smart plant cultivation service server promotes an environmentally friendly plant cultivation service by reducing the use of pesticides, fertilizers, and herbicides by providing information on the use of pesticides, fertilizers, and herbicides in advance, rather than presenting reactive control and treatment measures, by comprehensively analyzing plant growth factors caused by the plant growth status, weed status, and surrounding environment, and predicting pest and disease monitoring, the speed of spread of pests and diseases, and weed status. It also promotes plant growth by verifying the plant nutrient status and providing information on nutrients, water, and light intensity based on this to promote plant growth, while analyzing video images within the facility to analyze the weed status and the impact of the analyzed weed status on plant growth to ensure weeding is performed in advance, thereby preventing factors that inhibit plant growth in advance. A smart plant cultivation service system through pest and weed occurrence forecasting characterized by promoting an eco-friendly plant cultivation service by reducing the use of herbicides. Claim 2 delete Claim 3 delete Claim 4 delete Claim 5 A smart plant cultivation service system through pest and weed occurrence prediction in claim 1, wherein the data analysis unit comprises a collected data storage unit that stores the pest and treatment plan generated by the pest and treatment plan generation unit as analysis information. Claim 6 delete Claim 7 delete Claim 8 delete Claim 9 delete Claim 10 The smart terminal includes a smart terminal that predicts the occurrence of pests and weeds, the spread rate of pests and weeds, and plant nutritional status through AI-based analysis based on information collected through sensors acquiring information on the condition of cultivated plants in a facility and surrounding environment information, and information on plant condition and pests and diseases obtained through servers of public agricultural institutions, and suggests control and treatment methods based thereon. The smart terminal includes: a data collection unit that collects information collected by sensors and agricultural-related information through servers of public agricultural institutions; a data processing unit that preprocesses the cultivated plant condition information, surrounding environment information, and agricultural-related information collected by the data collection unit in frame units; a data analysis unit that classifies the data preprocessed by the data processing unit to generate a dataset, processes the generated dataset with a prediction model to generate prediction results, and generates control and treatment methods based on the generated prediction results; and an analysis information provision unit that provides the control and treatment methods generated by the data analysis unit as analysis information. The data analysis unit searches a database based on pest prediction results, pest spread rate prediction results, weed condition prediction results, or plant nutritional status prediction results to predict control and It includes a pest control and treatment plan generation unit that generates treatment plans, and the data analysis unit comprises: a dataset generation unit that constructs a dataset by labeling cultivated plant status information collected from a sensor; a model generation unit that generates a pest prediction model for pest monitoring, a pest spread speed prediction model for predicting the spread speed of pests, a weed state prediction model for predicting weed state, and a plant nutrient state prediction model for predicting plant nutrient state; a model training unit that trains the models generated by the model generation unit based on deep learning; and a model optimization unit that optimizes the pest prediction model, the pest spread speed prediction model, the weed state prediction model, and the plant nutrient state prediction model based on the training results of the model training unit.A prediction model application unit that applies a constructed dataset to a prediction model generated for the prediction of pest and disease monitoring and spread speed prediction of cultivated plants, weed status prediction, and plant nutritional status prediction, respectively; and a prediction result generation unit that generates prediction results based on the output results of the prediction model application unit and weather information, wherein the prediction model application unit includes: a pest and disease prediction model that performs pest and disease monitoring based on the state information of cultivated plants; a pest and disease spread speed prediction model that predicts the spread speed of pests and diseases based on the state information of cultivated plants; a weed status prediction model that predicts weed occurrence and weed spread speed based on the state information of cultivated plants; and a plant nutritional status prediction model that predicts the nutritional status of plants based on the state information of cultivated plants, wherein the control and treatment plan generation unit includes: a pest and disease determination unit that determines pests and diseases based on the pest and disease prediction results or the pest and disease spread speed prediction results; a control and treatment plan determination unit that derives control and treatment plans pre-registered in a database based on the pest and disease determination information determined by the pest and disease determination unit; and based on the weed status prediction results A weed state determination unit for determining weed state;It includes a weed treatment plan determination unit that derives weed treatment plans pre-registered in a database based on the weed status determined by the weed status determination unit; the control and treatment plan determination unit determines whether to use pesticides and fertilizers and the amount to use based on the pest and disease prediction results and the pest and disease spread speed prediction results; the weed treatment plan determination unit converts the weed status determined by the weed status determination unit into weights and applies the converted weights to a model analyzing the impact on pests and diseases and plant cultivation to determine whether to use herbicides and the amount to use; the weed status prediction model analyzes the presence and size of weeds by comparing weed images extracted from the state information of cultivated plants with reference weed images, and predicts weed occurrence and spread speed by quantifying these analysis results; the smart terminal comprehensively analyzes plant growth factors caused by plant growth status, weed status, and surrounding environment rather than presenting post-event control and treatment plans, thereby predicting pest and disease occurrence A smart plant cultivation service system through pest and weed occurrence forecasting, characterized by predicting the spread rate of pests and diseases and weed conditions to provide information on the use of pesticides, fertilizers, and herbicides in advance to reduce their use and promote environmentally friendly plant cultivation services; verifying plant nutritional status and providing information on nutrients, water, and light intensity for plant growth based on this to promote plant growth; analyzing weed conditions by analyzing video images within the facility and analyzing the impact of the analyzed weed conditions on plant growth to ensure weeding is performed in advance, thereby preventing factors that inhibit plant growth and reducing herbicide use to promote environmentally friendly plant cultivation services.

Images