KR20220071405A - Agricultural support system and method using big data of smart farm - Google Patents

Abstract

The present invention relates to a system and a method for supporting agriculture using smart farm big data. The system of the present invention includes: an interface unit collecting smart farm data including at least one of target crop growth information, environmental information, cultivation facility information, farming information, and management information; an integrated control unit analyzing the collected smart farm data and pre-provided data to generate an artificial intelligence model; and a management server including a management support unit. When smart farm management request information is received, the management support unit analyzes the received smart farm management request information in accordance with the generated artificial intelligence model to generate smart farm decision-making information. According to the present invention, smart farm big data including target crop growth information, environmental information, cultivation facility information, farming information, management information, or the like is collected, a cloud platform built with the collected smart farm big data and artificial intelligence technology are used, and thus agricultural decision making necessary for smart farm operation and management can be effectively supported.

Description

Agricultural support system and method using big data of smart farm

The present invention relates to an agricultural support system and method, and more particularly, agricultural support using big data of a smart farm that can support agricultural services using a cloud platform and artificial intelligence technology built on the basis of big data of a smart farm It relates to systems and methods.

In general, agriculture that produces food is an industry essential for human survival, but it is also one of the slowest industries to innovate because traditional agricultural technology is maintained. In particular, in domestic agriculture, management is deteriorating due to a decrease in the productive population, an aging population, labor costs, material costs, or cost problems of distribution channels. is gradually increasing.

In order to overcome this, various innovations in agriculture are being attempted. Among them, information and communication technology (ICT) is applied to agricultural production, processing, distribution, or consumption to remotely and automatically manage the growing environment of crops. In addition, smart farm, an agricultural system that can increase production efficiency, is currently attracting attention.

Smart farms according to the prior art are mostly technologies that simply open and close and control cultivation facilities based on environmental information such as temperature, humidity, illuminance, or carbon dioxide. There was a problem in that the environment had to be set or manipulated by the farmer himself.

In addition, in order to use smart farm technology, not only knowledge about agriculture but also ICT capabilities to understand and analyze data are required. There was a problem in that it was not easy to access the elderly farmers.

In addition, in the smart farm technology being developed so far, the development of information processing technology, information sharing technology, or information database technology required in various related fields has not been sufficiently developed.

Accordingly, in the conventional smart farm technology, which aims for convenience and productivity, it collects, analyzes and processes crop growth information by utilizing big data, cloud platform, and artificial intelligence technology of smart farms, and produces information on crop production. Development of an agricultural support system to optimally provide

Korean Patent Publication No. 2012-0028191 (published date: 2012.03.22) Korean Patent Publication No. 10-1871468 (Registration Date: 2018. 06. 20)

Therefore, the present invention collects big data of a smart farm including growth information of target crops, environmental information, cultivation facility information, farming information or management information, etc., and a cloud platform and artificial intelligence built with the collected big data of the smart farm. An object of the present invention is to provide an agricultural support system and method using big data of a smart farm that can support agricultural decision-making necessary for operation and management of smart farms using technology.

The problems of the present invention are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the following description.

Agricultural support system using big data of a smart farm according to an embodiment of the present invention for solving the above technical problem includes at least one of growth information of a target crop, environmental information, cultivation facility information, farming information, and management information When the interface unit for collecting smart farm data, the integrated control unit for generating an artificial intelligence model by analyzing the collected smart farm data and the data provided in advance, and the management request information of the smart farm are received, the received smart farm management request information and a management server including a management support unit for generating decision information of the smart farm by analyzing the generated artificial intelligence model.

The management support unit is the smart farm decision-making information including at least one of whether the target crop is pests or pests, manual information for measures against pests and pests, cultivation management information according to the growth status of each target crop, and failure and replacement information of the cultivation facility. may be generated and transmitted to the terminal device.

The smart farm management request information may be received through a terminal device according to a user's request or may be received from the smart farm device at predetermined intervals.

The smart farm device may include a sensor for measuring at least one of the growth of the target crop, the environment of the cultivation site, and the state of the cultivation facility, and a photographing device for photographing at least one of the growth of the target crop, the cultivation site and the cultivation facility. have.

When it is determined that the target crop is a pest, the cultivation facility or the smart farm device is determined to be malfunctioning, the management support unit may generate warning information notifying the occurrence and failure of the pest and transmit it to the terminal device.

The artificial intelligence model may be generated using fake data prepared to supplement the previously provided data.

On the other hand, the agricultural support method using big data of the smart farm performed by the management server that supports the agricultural support service through communication with the smart farm device and the terminal device according to the embodiment of the present invention is the growth information of a target crop, environmental information , collecting smart farm data including at least one of cultivation facility information, farming information, and management information, generating an artificial intelligence model by analyzing the collected smart farm data and data provided in advance, and requesting management of the smart farm and generating decision-making information of the smart farm by analyzing the received management request information of the smart farm according to the generated artificial intelligence model when the information is received.

The decision-making information of the smart farm may include at least one of whether the target crop is plagued by pests, manual information for measures against pests and pests, cultivation management information according to the growth state of each target crop, and failure and replacement information of the cultivation facility.

The smart farm management request information may be received through a terminal device according to a user's request or may be received from the smart farm device at predetermined intervals.

The smart farm device may include a sensor for measuring at least one of the growth of the target crop, the environment of the cultivation site, and the state of the cultivation facility, and a photographing device for photographing at least one of the growth of the target crop, the cultivation site and the cultivation facility. have.

The decision-making information of the smart farm may include warning information notifying the occurrence and failure of pests and pests when it is determined that the target crop is a pest, when the cultivation facility or the smart farm device is determined to be malfunctioning.

The artificial intelligence model may be generated using fake data prepared to supplement the previously provided data.

As described above, according to the agricultural support system and method using the big data of the smart farm according to the embodiment of the present invention, the big data of the smart farm including growth information, environmental information, cultivation facility information, farming information or management information of a target crop, etc. It has the advantage of effectively supporting agricultural decision-making required for smart farm operation management by collecting data and using the cloud platform and artificial intelligence technology built with the collected big data of the smart farm.

At this time, by using artificial intelligence functions such as machine learning or deep learning, automatic diagnosis of failures in cultivation facilities, analysis of the performance of measuring devices, and prediction-based maintenance platform are used to efficiently manage production and reduce maintenance costs It has the advantage of facilitating the integrated management and operation of the smart farm as it enables fast data analysis with the

In addition, by using the cloud platform, it is possible to diagnose and set the cultivation environment specialized for each farmhouse service, and it notifies the contents of cultivation management, yield, harvest time prediction, risk detection for unusual situations, failure diagnosis or prediction, etc. to be reflected in decisions.

It acquires information on target crops in real time from devices installed in multiple smart farms, acquires numerous data from terminal devices owned by agricultural workers, household smart farms, and users related to crop cultivation to build big data in the cloud environment, It has the advantage of being able to properly operate a smart farm based on built-in big data and artificial intelligence algorithms.

In addition, the predictive maintenance system improves reliability, prevents secondary accidents in advance, reduces maintenance costs, and enables systematic management and supervision of farms operated by various smart farm systems, thereby improving farming convenience and It has the advantage of improving productivity.

In addition, in the management and operation of the smart farm, it is possible to estimate the replacement time of the cultivation facility through risk detection, failure diagnosis, and failure prediction according to the individual characteristics of various smart farms, and to prevent sudden accidents. has a very effective advantage.

As such, young or returning farmers with little knowledge of agriculture, elderly farmers who have agricultural knowledge but are not familiar with ICT, and novice farmers who lack ICT competency and ability to understand and analyze data, use this system. This has the advantage of being able to create the same effect as a farmer with farming experience.

In addition, the development and maintenance of smart farms is easy, effective in operation and management, and it is possible to build a reasonable agricultural management infrastructure suitable for the artificial intelligence of the 4th industrial revolution in the future.

The effect according to the present invention is not limited by the contents exemplified above, and more various effects are included in the present specification.

1 is a configuration diagram schematically showing an agricultural support system using big data of a smart farm according to an embodiment of the present invention.
2 is a detailed configuration diagram of an agricultural support system using big data of a smart farm according to an embodiment of the present invention.
3 is a block diagram of a neural network model for producing fake image data according to an embodiment of the present invention.
4 is an exemplary diagram showing a neural network learning process for pests according to an embodiment of the present invention.
5 is an operation flowchart illustrating an agricultural support process using big data of a smart farm according to an embodiment of the present invention.
6 is an exemplary diagram illustrating a process of using an agricultural support service through a terminal device.

Advantages and features of the present invention, and a method for achieving them will become apparent with reference to the embodiments described below in detail in conjunction with the accompanying drawings.

However, the present invention is not limited to the embodiments disclosed below, but will be implemented in various different forms.

In the present specification, the present embodiment is provided so that the disclosure of the present invention is complete, and to completely inform those of ordinary skill in the art to which the present invention belongs, the scope of the invention.

And the invention is only defined by the scope of the claims.

Accordingly, in some embodiments, well-known components, well-known operations, and well-known techniques have not been specifically described to avoid obscuring the present invention.

In addition, the same reference numerals refer to the same elements throughout the specification, and the terms used (referred to) in this specification are for the purpose of describing the embodiments, and are not intended to limit the present invention.

In this specification, the singular also includes the plural unless specifically stated otherwise in the phrase, and elements and operations referred to as 'comprising (or having)' do not exclude the presence or addition of one or more other elements and operations. .

Unless otherwise defined, all terms (including technical and scientific terms) used herein may be used with the meaning commonly understood by those of ordinary skill in the art to which the present invention belongs.

In addition, terms defined in a commonly used dictionary are not ideally or excessively interpreted unless they are defined.

Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings.

1 is a configuration diagram schematically showing an agricultural support system using big data of a smart farm according to an embodiment of the present invention, and FIG. 2 is an agricultural support system using big data of a smart farm according to an embodiment of the present invention. A detailed configuration diagram is shown.

1 and 2 , the agricultural support system 1 using big data of the smart farm includes a smart farm device 100 , a terminal device 200 , a management server 300 , and a database 400 . is composed by

The smart farm device 100 is composed of at least one, and collects smart farm data such as growth information, environmental information, cultivation facility information, farming information or management information of the target crop generated in the process of growing the target crop to be described below. It may be transmitted to the management server 300 or the database 400 . In this case, the smart farm data may be formed of structured data or unstructured data such as images.

The growth information of the target crop is information indicating the growth state of the target crop, and may include, for example, the leaf length, leaf width, number of leaves, stem diameter, height of a flower room, number of fruits, number of flowers, growth length or overweight of the target crop. can

The environmental information may include information about the internal environment, external environment, soil or water environment of the smart farm. At this time, information on the internal environment of the smart farm includes information on temperature, humidity, carbon dioxide, light or insolation, etc. inside the cultivation site, and information on the external environment of the smart farm includes temperature, wind direction, wind speed, It may include information about rainfall or solar radiation. Information on soil conditions includes information on soil moisture, ground temperature, irrigation amount, EC or pH, etc., and hydroponic information includes water supply amount, supply pH, supply EC, soil moisture, geo temperature, moisture content, drainage amount, drainage pH or drainage EC, etc. may include information about

Cultivation facility information largely includes information on cultivation facilities and control information of cultivation facilities. Information on cultivation facilities includes basic information including type of cultivation facility, cultivation type, cultivation area, location, medium type, etc., and integrated controller. (manufacturer, type, model, location, quantity), nutrient solution/irrigation device (manufacturer, type, model, location, quantity), air conditioner (manufacturer, type, model, location, quantity), energy information (energy type, usage, management method), and sensor information indicating the type, location, and normal range of the sensor, and actuator information indicating the type, quantity, location, and automatic presence of the actuator. The control information of the cultivation facility includes, for example, skylights, side windows, thermal curtains, shading curtains, floating fans, auxiliary lights, valves, exhaust fans, ventilation holes, irrigation, pumping water, moisture/management devices, heaters, air conditioners, fumigators, water film devices, or carbon dioxide. It may include information indicating whether or not the cultivation facility such as the generator is operating, operating time, operating status, setting status, operating temperature, and the like.

Agricultural information may include information on planting date and time, crop/cultivar, planting density, red leaf date and time, red leaf amount, wetting/wetting time, proper seed cutting, and excess amount. Information on unit price, production cost, etc. may be included.

In addition, of course, the smart farm device 100 may collect various smart farm data according to various environments of farmhouses or users.

The smart farm device 100 can measure the temperature, humidity, light quantity, carbon dioxide or soil condition (PH, EC, temperature, humidity, weight of the medium) of the cultivation site in real time, or detect the condition of the cultivation facility. The sensor 110 and the photographing device 130 for photographing a cultivation site or a cultivation facility may be included. In addition, the smart farm device 100 may further include a sensor 110 capable of sensing and controlling an amount of a material required for cultivation of a target crop, for example, a nutrient solution, water, a functional useful material, and the like.

And the smart farm data includes information measured by the sensor 110 or the photographing device 130, information for identifying the sensor 110 or the photographing device 130, and indicating the state of the sensor 110 or the photographing device 130. Information, information indicating the position of the sensor 110 or the photographing device 130 , and information recorded with the measurement time of the sensor 110 or the photographing time of the photographing apparatus 130 may be further included.

The photographing device 130 may be a camera installed in the cultivation site, or a mobile photographing apparatus such as a drone if there is no camera. When smart farm data is collected with a mobile photographing device such as a drone, basic information for collection history management can be additionally generated. The additional information may be location information of a place currently using the photographing device, a unique number for management, and information determined to be additionally managed according to a user's request. Examples of the photographing location may be various locations such as a smart farm farm or a plant factory, and location information of the mobile photographing device may be obtained using an algorithm such as a location based service (LBS).

As an algorithm for securing location information of a mobile photographing device, after location information is collected by trilateration and fingerprinting, each degree of similarity can be determined and final location information can be first corrected. In this case, the cosine similarity can be used because the tolerance for the similarity criterion uses the value specified in the initial value setting, and the similarity must have the directionality of the vector as a characteristic. And after the first correction of the location information is performed, the amount of learning (cumulative) data of the initial value setting is checked, and if the amount of learning data does not meet the standard, after collecting the location information by trilateration and fingerprint method, each The process of finally correcting the location information by determining the degree of similarity may be repeatedly performed. If the amount of training data is sufficiently secured (the reference value of the initial value setting), secondary correction of the location information is performed, fake data is generated using a neural network (GAN), similarity determination and initial value By determining the specified accuracy of the setting, the location information may be updated or the above process may be repeated. At this time, the reason for using the neural network is that the amount of data of the existing location information to be used for learning is too small, in order to obtain more accurate location information by securing a large amount of learning data, and the result of the data of location information learned with sufficient data If a certain degree of reliability is accepted, this method can be selectively used.

And, in general, the fingerprint LBS technology uses Wi-Fi, a beacon (Bluetooth) module, etc., but in an embodiment of the present invention, a beacon module having lower power and lower price than a Wi-Fi module may be used. In addition, since the mobile photographing device has its own camera, it can communicate with the management server 300 to provide information on pests and pests and measures in real time.

The terminal device 200 is a personal computer (Personal Computer, PC), smart phone (Smart Phone), tablet (Tablet) PC, personal digital assistant (PDA) or memory means such as a web pad (Web Pad) It may be a terminal equipped with a arithmetic capability and equipped with a microprocessor.

The terminal device 200 may be a terminal possessed by a user who cultivates or manages a target crop, a user to acquire pest information, a user to utilize an action method for pests or pests, or to receive support in the form of consulting, in addition to a home smart farm user Of course, it may be made of a terminal, a robot, or other mobile device.

In particular, the terminal device 200 may be installed with an agricultural support service application or program that generates management request information of the smart farm according to a user's request and transmits it to the management server 300 . For example, when the user suspects that pests and pests have occurred in the target crop, the user may execute an agricultural support service application or program of the terminal device 200 to photograph the target crop or record the target crop. In addition, the captured image or recorded data of the target crop may be transmitted to the management server 300 as management request information of the smart farm.

In addition, the smart farm device 100 generates management request information of the smart farm under the control of the terminal device 200 and transmits it to the management server 300 , or transmits the management request information of the smart farm to the smart farm device 100 at a predetermined period in the smart farm device 100 . Management request information may be automatically generated and transmitted to the management server 300 . For example, a camera installed at a cultivation site or a cultivation facility captures an image of a target crop or a cultivation facility at predetermined intervals and transmits it to the management server 300 to determine whether the target crop is pest or whether the cultivation facility is malfunctioning. let it be

The management server 300 may exchange various information with the smart farm device 100 , the terminal device 200 , the agricultural institution server (not shown), and the database 400 through the communication network 10 .

Here, the communication network 10 includes a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), the Internet, and 3G/4G/ 5G (generation) mobile communication network, WiFi (Wi-Fi), WIBRO (Wireless Broadband Internet) or LTE (Long Term Evolution) may include various data communication networks, including LTE (Long Term Evolution), etc. It doesn't matter if you use

In more detail, the management server 300 may include an interface unit 310 , an integrated control unit 330 , and a management support unit 350 .

The interface unit 310 receives smart farm data, such as growth information, environmental information, cultivation facility information, farming information or management information, transmitted from the smart farm device 100, the terminal device 200, or an agricultural institution server, etc. to the database ( 400) can be stored and collected.

The interface unit 310 may receive or transmit smart farm data from the outside in an OpenAPI or Software Development Kit (SDK) method. For example, data sets provided by institutions or organizations such as the Rural Development Administration can be collected and used as data for artificial intelligence learning or prediction, and external decision-making is made using the collected, analyzed, and learned artificial intelligence model. A data set in the form of a provided or raw data for support may be provided externally.

OpenAPI can be provided through wired/wireless-based communication, and SDK can be installed in the form of a module in the case of facility S/W operation and development to support the operation of cultivation facilities. For example, by providing pest management, agricultural information, various management information or control information, etc., farmers or returning farmers with little knowledge of agriculture, elderly farmers with agricultural knowledge but not familiar with ICT, and ICT competency and data can be understood and analyzed. It can help novice farmers who lack the ability to do so.

In an embodiment of the present invention, the interface unit 310 is described as being operated inside the management server 300 , but it is also possible to operate outside the management server 300 .

The integrated control unit 330 may generate an artificial intelligence model by analyzing the collected smart farm data and previous data stored in the database 400 .

More specifically, the integrated control unit 330 may generate an artificial intelligence model by performing analysis and processing on the collected smart farm data and the big data stored in the database 400 because the artificial intelligence function is implemented. . Here, artificial intelligence is a field of computer engineering and information technology that studies how computers can do the thinking, learning, or self-development that human intelligence can do. to make it possible Artificial intelligence is a technology that realizes human learning, reasoning, perception, and language ability as a computer program, and may include machine learning or deep running algorithms, etc. It is a field that teaches how to learn on its own. It is a mathematical algorithm or probabilistic model that learns from data such as logic or standardized rules, and is a method of making assumptions or predictions based on similar data sets.

Machine learning is largely divided into supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning. In supervised learning, a desired answer called a label is included in the training data in a data set, and a learning model is created using this. In unsupervised learning, a prediction result can be obtained by creating a learning model with a data set that does not have correct answer data and applying it. Semi-supervised learning can be trained using a data set with only partial labels. Reinforcement learning can be learned as a process of obtaining rewards or penalties as a result of performing an action.

Deep learning is a field of machine learning, similar to machine learning, but it refers to a technology for predicting or classifying new data sets as a neural network that learns by creating and using a structure called an artificial neural network by the structure of neurons created by referring to the human brain. .

In this way, by collecting almost all data provided by the smart farm device 100, the terminal device 200, or the agricultural institution server, etc., the accuracy of AI learning is increased, and furthermore, an effective decision-making support platform for those involved in agriculture and crop cultivation is provided. to be able to provide

When the management support unit 350 receives the smart farm management request information from the smart farm device 100 or the terminal device 200 , the received smart farm management request information is integrated with the learned artificial intelligence generated by the integrated control unit 330 . It is possible to generate decision-making information for smart farms by analyzing them according to the model. In addition, the management support unit 350 transmits the generated smart farm decision-making information to the smart farm device 100 or the terminal device 200 through the interface unit 310 to receive information about a user who cultivates or manages a target crop, pests and diseases. Users who want to acquire it, use measures against pests and diseases, or users who want to receive support in the form of consulting, etc. can check it.

The management support unit 350 implements an artificial intelligence model that can be updated in real time and maintained in the latest state. When the management request information of the smart farm is received through the interface unit 310, the photo included in the management request information of the smart farm Or, it is possible to determine whether the target crop is diseased or not by analyzing the video with an artificial intelligence model.

The management support unit 350 determines whether or not the cultivation facility is faulty or replacement time, whether the smart farm device 100 in the cultivation facility is faulty or replacement time, and is a smart farm for a manual that can notify or take action. Decision-making information can be generated. For example, when the target crop is determined to be a pest, or the cultivation facility or the smart farm device 100 is determined to be malfunctioning, the management support unit 350 provides warning information such as a message, vibration or sound form informing of the occurrence and failure of pests and pests. By transmitting to the terminal device 200 , the user of the terminal device 200 can quickly take action against pests or failures.

Smart farm management information is diverse for each crop and there are many types, so it is necessary to collect scattered data. It can be provided in the form of an OpenAPI or SDK that can be used.

In this way, the management support unit 350 may generate decision-making information of a smart farm that supports growth information for supporting crop recommendation models and crop-specific growth information based on data such as soil information and environmental information of the current region. , pests can be judged by a neural network (CNN series) model trained with existing open data. It can also be predicted using the predictive model provided by the Promotion Agency. In addition, the determination or prediction of the failure of the smart farm device 100 or the cultivation facility can be determined and predicted by the processing method of time series data of an artificial intelligence neural network (RNN series), and the latest model based on the data obtained in real time can be updated. As described above, the process of determining and predicting pests using the neural network model will be described in more detail with reference to FIGS. 3 and 4 below.

In the case of device failures, pests, or management information, there are parts that are actually learned by artificial intelligence, and it is possible to predict and determine failures for each symptom based on real-time data. may be provided in the form of a data set of

In this way, the management server 300 increases the accuracy of artificial intelligence learning by collecting numerous data provided by the smart farm device 100, the terminal device 200, or an agricultural institution server, and furthermore, it is possible to make effective decisions about agriculture. can support you to do so.

On the other hand, in an embodiment of the present invention, the functions and operations of the integrated control unit 330 and the management support unit 350 are divided and described to be implemented in separate components, but the present invention is not limited thereto and the integrated control unit 330 and the management support unit The functions and operations of 350 may be performed in one component.

The database 400 may store smart farm data such as growth information, environmental information, cultivation facility information, farming information or management information of a target crop received from the smart farm device 100 or the terminal device 200, and agricultural Big data received from an institution server or the like may be stored. In addition, since the database 400 has different crop growth information for each target crop type (eg, tomato, cucumber, eggplant, strawberry, etc.), recipe information for managing it or supporting it through external consulting may be stored.

In this case, the database 400 may be installed inside the management server 300 , or may be installed separately from the management server 300 .

Hereinafter, neural network learning for judging and predicting pests based on images will be described.

A large amount of image data is required to train a neural network to determine and predict pests based on images, but it is not easy to collect most of the data. Accordingly, in one embodiment of the present invention, a neural network model is used as an artificial intelligence model, but image data insufficient for neural network learning is replaced with fake image data prepared in advance, and an unsupervised learning neural network model, a Generative Adversarial (GAN) model, is used. Nets) model is applied to neural network learning to increase the accuracy of the judgment and prediction of pests.

More specifically, in order to increase the accuracy of the prediction and judgment of pests and pests, the advantages of VAE (Variational Auto Encoder) and the characteristics of the GAN model are used. The GAN, which is often judged by visualizing the results because the criteria are not clear, is placed at the back so that a high-accuracy neural network model can be used.

3 is a block diagram of a neural network model for producing fake image data according to an embodiment of the present invention.

Referring to FIG. 3 , a neural network model for producing fake image data may largely include a filter unit 510 , a VAE unit 520 , a CNN unit 530 , and a GAN unit 540 , and an input Image data obtained previously can be used for this. The input image data is subjected to a pre-processing process such as noise removal through the filter unit 510 to correct various distortions, and an Image Augmentation (Crop, Rotate, Flip, Translate, Resize, etc.) technique can be used. The distortion-corrected image data or image data generated by the Image Augmentation technique is input to the VAE unit 520 composed of multiple inputs, and is processed by the encoder (Encoder, 521), the Latent Space (523) and the decoder (Decoder, 525). By passing through , a plurality of latent feature vector data may be output.

Here, Image Augmentation (image data modulation) technique is a method of increasing the accuracy of prediction and judgment of pests and pests with artificial intelligence. In this outbreak, it is to solve the data shortage phenomenon (see Fig. 3 and Fig. 4) necessary for artificial intelligence model learning, and when sufficient data is secured, the data obtained in the pre-processing step is used, or for higher accuracy learning Other methods proposed in the present invention may be used in combination.

The simplest modulation methods are crop, rotate, and flip, and if you think the center part of the image is important, you can crop the pixels on the edges. At this time, it is also rotated at a random angle. It is also possible to flip left and right and up and down. When using a patch, which is a small cropped part of an image, if you cut out 250x250 patches from a 300x300 image, you can get 5 images with the same label just by cropping them from the four corners and from the center. When it is necessary to achieve better performance with only limited data, a method of randomly pruning a patch or pruning with a predetermined rule is a method frequently used. On the other hand, while maintaining the size of the image, it is sometimes translated by a few pixels up, down, left and right. In this case, an empty space is created, which can be filled with zero (zero filling), filled with the nearest pixel (nearest neighbor), or filled with a pushed part (rolling). In addition, in case the size of the image is different, it is possible to adjust the size to the same size and perform training through batch rescale. Taking pedestrian recognition as an example, if the size is different, such as some people are within 32x56 pixels, some are within 41x78 pixels, etc., the size can be adjusted to fit within 32x64 pixels by collectively increasing and decreasing a predetermined range. In the case of an image where the aspect ratio is important, you can align the short side and cut it out, or align the long side and fill in the blanks. In addition, zooming is also possible.

A plurality of potential feature vector data passes through the CNN unit 530 and is used as input data of P(x) 541 of the neural network GAN model, and random vector data can be used as input data of G(x) 543. have. In the D(x) 545 of the GAN unit 540, a loss function is obtained through the mutual difference between the P(x) 541 and the G(x) 543, and a backproper is obtained using the obtained loss function. Fake image data and real image data are corrected through a backpropagation method. By repeating this process, more precise prediction is possible, and the output data of D(x) 545 can be used for final training, testing, or model validation.

In this case, the CNN unit 530 may configure multiple neural network layers, and other neural networks of supervised learning, unsupervised learning, or semi-supervised learning may be configured in the layer to configure various learning data.

4 is an exemplary diagram showing a neural network learning process for pests and pests according to an embodiment of the present invention. As in FIG. 4, the types of crops and pest types (leaf, fruit, insects, etc.) are selected (S400), It may be determined whether or not to learn the neural network (S410).

When the neural network learning is performed in step 410 (S410-Y), it can be checked whether to use the existing pest image data (S420). If the amount of the existing pest image data is insufficient (S420-N), the The fake pest image data may be received from the fake pest image database created in this way, and the neural network may be trained (S430) and used to generate a neural network model (S440). When using existing pest image data (S420-Y), neural network learning from the pest database is performed (S450), a neural network model is created (S440) and stored (S460), and the stored neural network model is later updated to the pest and pest neural network model. Make it usable for prediction and judgment.

On the other hand, if neural network learning is not performed in step 410 (S410-N), a neural network model suitable for the crop type and pest type selected from the previously trained pest and pest neural network model database is used (S470) to perform pest prediction and judgment. can be (S480).

Hereinafter, an agricultural support method using big data of a smart farm according to an embodiment of the present invention will be described.

5 is a flowchart illustrating an agricultural support process using big data of a smart farm according to an embodiment of the present invention.

Referring to FIG. 5 , the interface unit of the management server may collect smart farm data including at least one of growth information, environment information, cultivation facility information, farming information, and management information of a target crop ( S500 ). In this case, the smart farm data may consist of structured data or unstructured data such as images, and may be collected and received from a smart farm device, a terminal device, or an agricultural institution server.

In this case, the interface unit may receive or transmit the smart farm data from the outside in an OpenAPI or SDK (Software Development Kit) method. For example, data sets provided by organizations or organizations such as the Agricultural Promotion Administration can be collected and used as data for artificial intelligence learning or prediction, and external decision-making is made using the collected, analyzed, and learned artificial intelligence model. A data set in the form of a provided or raw data for support may be provided externally. OpenAPI can be provided through wired/wireless-based communication, and SDK can be installed in the form of a module in the case of facility S/W operation and development to support the operation of cultivation facilities. For example, by providing pest management, agricultural information, various management information or control information, etc., farmers or returning farmers with little knowledge of agriculture, elderly farmers with agricultural knowledge but not familiar with ICT, and ICT competency and data can be understood and analyzed. It can help novice farmers who lack the ability to do so.

Here, the smart farm device is a sensor that can measure the temperature, humidity, light quantity, carbon dioxide or soil condition (PH, EC, temperature, humidity, weight of the medium) of the cultivation site in real time or detect the condition of the cultivation facility , a photographing device for photographing a cultivation site or a cultivation facility, and the like. In addition, the smart farm device may further include a sensor capable of sensing and controlling the amount of a material required for cultivation of a target crop, for example, a nutrient solution, water, a functional useful material, and the like.

In addition, smart farm data includes information measured by a sensor or a photographing device, information for identifying a sensor or photographing device, information indicating the state of a sensor or photographing device, information indicating the location of a sensor or photographing device, measuring time or photographing of the sensor Information such as recording time of the device may be further included.

The photographing apparatus may be a camera installed in the cultivation site, or a mobile photographing apparatus such as a drone if there is no camera. When smart farm data is collected with a mobile photographing device such as a drone, basic information for collection history management can be additionally generated. The additional information may be location information of a place currently using the photographing device, a unique number for management, and information determined to be additionally managed according to a user's request. Examples of the photographing location may be various locations such as a smart farm farm or a plant factory, and location information of the mobile photographing device may be obtained using an algorithm such as a location based service (LBS).

Next, the integrated control unit of the management server may generate an artificial intelligence model by analyzing the collected smart farm data and data provided in advance (S510).

The integrated control unit has an artificial intelligence function implemented, so it can create an artificial intelligence model by analyzing and processing the collected smart farm data and big data stored in the database. Here, artificial intelligence is a field of computer engineering and information technology that studies how computers can do the thinking, learning, or self-development that human intelligence can do. to make it possible Artificial intelligence is a technology that realizes human learning, reasoning, perception, and language ability as a computer program, and may include machine learning or deep running algorithms, etc. It is a field that teaches how to learn on its own. It is a mathematical algorithm or probabilistic model that learns from data such as logic or standardized rules, and is a method of making assumptions or predictions based on similar data sets.

Machine learning is largely divided into supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning. In supervised learning, a desired answer called a label is included in the training data in a data set, and a learning model is created using this. In unsupervised learning, a prediction result can be obtained by creating a learning model with a data set that does not have correct answer data and applying it. Semi-supervised learning can be trained using a data set with only partial labels. Reinforcement learning can be learned as a process of obtaining rewards or penalties as a result of performing an action.

Deep learning is a field of machine learning, similar to machine learning, but it refers to a technology for predicting or classifying new data sets as a neural network that learns by creating and using a structure called an artificial neural network by the structure of neurons created by referring to the human brain. .

In this way, it is possible to increase the accuracy of artificial intelligence learning by collecting almost all data provided from smart farm devices, terminal devices, or agricultural institution servers, and furthermore, to provide an effective decision support platform to those involved in agriculture and crop cultivation.

And when it is determined whether the management request information of the smart farm is received (S520) and the management request information of the smart farm is received (S520-Y), the management support unit of the management server generates the received management request information of the smart farm with artificial intelligence It is possible to generate decision-making information of the smart farm by analyzing it according to the model (S530).

In more detail, when the management support unit receives the smart farm management request information from the smart farm device or terminal device, it analyzes the received smart farm management request information according to the learned artificial intelligence model generated by the integrated control unit, and analyzes the smart farm management request information. Decision-making information can be generated. In addition, the management support unit transmits the generated smart farm decision-making information to the smart farm device or terminal device through the interface unit to cultivate or manage the target crop, use the user to acquire pest information, Allow users, etc. to check to receive support in the form of consulting.

The management support department implements an artificial intelligence model that can be updated in real time to keep it up to date. When the management request information of the smart farm is received through the interface unit, the photo or video included in the management request information of the smart farm is converted into an artificial intelligence model. By analyzing it, it is possible to determine whether the target crop is pest or not.

The management support department determines whether the cultivation facility is faulty or when to replace it, whether or not the smart farm device in the cultivation facility is faulty or when to replace it, and generates smart farm decision-making information for a manual that can notify or take action. can For example, when the target crop is determined to be a pest, or the cultivation facility or the smart farm device 100 is determined to be malfunctioning, the management support unit 350 provides warning information such as a message, vibration or sound form informing of the occurrence and failure of pests and pests. By transmitting to the terminal device 200 , the user of the terminal device 200 can quickly take action against pests or failures.

In this way, the management support unit can create a smart farm decision-making information that supports growth information to support a crop recommendation model and crop-specific growth information based on data such as current local soil information and environmental information. Judgment can be made with a neural network (CNN series) model trained with existing public data, and the prediction of pests and pests is provided by data-based trend prediction, statistical artificial intelligence method, or the Agricultural Promotion Administration when sufficient data is accumulated during operation. It can also be predicted using a predictive model that is In addition, judgment or prediction of the failure of smart farm devices or cultivation facilities can be determined and predicted by the processing method of time series data of artificial intelligence neural networks (RNN series), and the latest model can be updated based on the data obtained in real time. can In the case of device failures, pests, or management information, there are parts that are actually learned by artificial intelligence, and it is possible to predict and determine failures for each symptom based on real-time data. may be provided in the form of a data set of

In this way, the management server may collect a large number of data provided from a smart farm device, a terminal device, or an agricultural institution server, etc. to increase the accuracy of AI learning and further support effective decision-making on agriculture.

6 is an exemplary view illustrating a process of using an agricultural support service through a terminal device. Referring to FIG. 6 , a series of processes in which a user of the terminal device uses the agricultural support service is shown.

When the agricultural support service application or program is installed in the user's terminal device and the agricultural support service application or program is executed (S600), it may be determined whether the user is authenticated (S610). At this time, when the authentication is completed (S610-Y), a member registration procedure can be performed (S620).

In addition, information for requesting management of the smart farm may be transmitted to the management server by uploading a photo or video of the target crop so as to check damage by pests and diseases of the target crop (S630). Then, the management server can receive the smart farm decision-making information in the form of documents, voices, videos, etc., which is manual information on the process of detecting and confirming pest information using the artificial intelligence model and how to take action against it (S640) ).

If the decision-making information of the smart farm is insufficient, the person in charge providing the agricultural support service may additionally support it.

Those of ordinary skill in the art to which the present invention pertains will understand that the present invention may be embodied in other specific forms without changing the technical spirit or essential features thereof. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive. The scope of the present invention is indicated by the following claims rather than the above detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalent concepts should be interpreted as being included in the scope of the present invention. do.

1: Agricultural support system using big data of smart farm
100: smart farm device
200: terminal device
300: management server
310: interface unit
330: integrated control
350: Administrative Support Department
400: database

Claims (12)

an interface unit for collecting smart farm data including at least one of target crop growth information, environmental information, cultivation facility information, farming information, and management information;
an integrated control unit for generating an artificial intelligence model by analyzing the collected smart farm data and data provided in advance; and
a management server including a management support unit for generating decision-making information of the smart farm by analyzing the received management request information of the smart farm according to the generated artificial intelligence model when the management request information of the smart farm is received;
Agricultural support system using big data of smart farms, including
In claim 1,
The management support department,
A terminal device by generating decision-making information of the smart farm including at least one of whether the target crop is pests or not, manual information for measures against pests and pests, cultivation management information according to the growth state of each target crop, and failure and replacement information of the cultivation facility Agricultural support system using big data of smart farm transmitted to
In claim 1,
Management request information of the smart farm,
Agricultural support system using big data of a smart farm that is received through a terminal device according to a user's request or is received every predetermined period from a smart farm device.
In claim 3,
The smart farm device,
Big data of a smart farm comprising a sensor for measuring at least one of the growth of the target crop, the environment of the cultivation site, and the state of the cultivation facility, and a photographing device for photographing at least one of the growth of the target crop, the cultivation site and the cultivation facility Agricultural support system used.
In claim 4,
The management support department,
When it is determined that the target crop is a pest, when the cultivation facility or the smart farm device is determined to be malfunctioning, warning information notifying the occurrence and failure of the pest is generated and transmitted to the terminal device using big data of the smart farm Agricultural support system.
In claim 1,
The artificial intelligence model is
Agricultural support system using big data of smart farms generated using fake data produced to supplement the previously provided data.
In an agricultural support method using big data of a smart farm performed by a management server that supports an agricultural support service through communication with a smart farm device and a terminal device,
collecting smart farm data including at least one of growth information, environmental information, cultivation facility information, farming information, and management information of a target crop;
generating an artificial intelligence model by analyzing the collected smart farm data and data provided in advance; and
generating smart farm decision-making information by analyzing the received smart farm management request information according to the generated artificial intelligence model when the smart farm management request information is received;
Agricultural support method using big data of smart farms, including.
In claim 7,
The smart farm decision-making information,
Agricultural support method using big data of smart farms, which includes at least one of whether the target crop is pests or not, manual information for measures against pests and pests, cultivation management information according to the growth status of each target crop, and failure and replacement information of cultivation facilities.
In claim 7,
Management request information of the smart farm,
Agricultural support method using big data of a smart farm that is received through a terminal device according to a user's request or is received every predetermined period from a smart farm device.
10. In claim 9,
The smart farm device,
Big data of a smart farm comprising a sensor for measuring at least one of the growth of the target crop, the environment of the cultivation site, and the state of the cultivation facility, and a photographing device for photographing at least one of the growth of the target crop, the cultivation site and the cultivation facility Agricultural support methods used.
11. In claim 10,
The smart farm decision-making information,
Agricultural support method using big data of a smart farm that includes warning information for notifying the occurrence of pests and failures when the target crop is determined to be a pest and when the cultivation facility or the smart farm device is determined to be malfunctioning.
In claim 7,
The artificial intelligence model is
Agricultural support method using big data of smart farms generated using fake data produced to supplement the previously provided data.

Images