KR102300799B1 - Crop state judgment apparatus and crop cultivation information service platform including the same - Google Patents

Abstract

Embodiments of the present invention relates to a crop cultivation information service platform comprising a processor and a memory for storing a program executed by the processor, wherein the processor creates a flight path of an unmanned aerial vehicle according to a lot formed in consideration of preset topographic information and take-off and landing points, receives and analyzes crop images acquired by unmanned aerial vehicles to obtain crop information according to the flight path, and in consideration of crop information including the type of crop and the condition of the crop, generates a crop map that displays the harvest status and disaster status for each cultivation area according to the flight path. A main purpose of the present invention is to strengthen the predictive power of supply and demand by analyzing the images acquired through the unmanned aerial vehicle and displaying the types of crops by area on the crop map to quickly and rapidly grasp the situation of shipment fluctuations in the cultivation area and unusual weather conditions.

Description

Crop state judgment apparatus and crop cultivation information service platform including the same

The present invention relates to a crop condition determination device and a crop cultivation information service platform including the same, and in particular, by using an image obtained through a drone, the type of crop for each cultivation area of a crop is determined through the crop condition determination device, and the crop is selected. It relates to a crop cultivation information service platform for generating a displayed crop map.

The content described in this section merely provides background information for the present embodiment and does not constitute the prior art.

Precise agricultural observations are needed as variability in crop supply and demand increases due to extreme weather, changes in cultivated land use, crop renewal, changes in farming methods, and scale-up of cultivation due to global warming. Statistics to determine the current status of crops and trees in agricultural and forestry land are basic information for predicting the supply and demand of agricultural products and forest products and establishing future supply and demand plans. However, the existing methods are limited to the method of constructing spatial information based on building technology or survey technology, and there is an insufficient limit as data for understanding the vegetation in the actual field. That is, there is a limit in that future demand forecasts are also inaccurate because accurate information on current vegetation identification cannot be provided according to the existing method.

In the past, institutions such as the Rural Economic Research Institute and the National Statistical Office provided information on vegetation in farmland and mountainous areas. This is because the basic area of the land grown on the basis of demand forecasting itself was wrong.

Embodiments of the present invention have been devised to solve the above problems, and by analyzing images acquired through an unmanned aerial vehicle, the types of crops by area are displayed on the crop map to quickly and quickly respond to changes in the shipment of cultivated areas and unusual weather conditions. The main purpose of the invention is to strengthen the predictive power of supply and demand by understanding the situation.

Other objects not specified in the present invention may be additionally considered within the scope that can be easily inferred from the following detailed description and effects thereof.

According to one aspect of this embodiment, the present invention is a crop cultivation information service platform comprising a processor and a memory for storing a program executed by the processor, the processor, the preset topographic information and take-off point and landing point Creates a flight path of the unmanned aerial vehicle according to the land formed in consideration of the above, receives and analyzes the crop image obtained by the unmanned aerial vehicle to obtain crop information according to the flight path, and includes the type of crop and the state of the crop We propose a crop cultivation information service platform, characterized in that the crop map is generated in which the harvest status and disaster status for each cultivation area according to the flight path are displayed in consideration of the crop information.

According to another embodiment of the present invention, the present invention provides a crop status determination apparatus including a processor and a memory for storing a program executed by the processor, wherein the processor includes a plurality of classification models according to the type of crop and the It learns a plurality of state models according to the state of crops, receives crop images acquired by an unmanned aerial vehicle moving along a flight path, and analyzes them through a convolutional neural network (CNN) to classify crops by crop type. And it proposes a crop status determination device characterized in that the acquisition of the crop information including the status of the crop.

As described above, according to the embodiments of the present invention, since the present invention enables object recognition of the crop image thus obtained through an artificial neural network, the effect of more accurately grasping the status of crops by lot compared to the existing technology is effective. have.

In addition. According to the embodiments of the present invention, the present invention determines the crop type according to the crop image, determines the crop status according to the crop type, and generates and provides a crop map displaying the crop status according to the crop type for each area. Accordingly, it is effective to understand the current status of crops in detail.

Even if it is an effect not explicitly mentioned herein, the effects described in the following specification expected by the technical features of the present invention and their potential effects are treated as if they were described in the specification of the present invention.

1 is an exemplary diagram illustrating a crop cultivation information service using a crop cultivation information service platform according to an embodiment of the present invention.
2 is a block diagram illustrating and describing a computing environment including a computing device suitable for use in example embodiments.
3 is a view sequentially illustrating a process of obtaining crop cultivation information according to an embodiment of the present invention.
4 is a diagram of extracting a representative point of lots according to an embodiment of the present invention, and generating a flight trajectory according to the extracted representative point of a lot.
5 is a diagram illustrating a concept of performing image learning through AI and recognizing a type of crop through this according to an embodiment of the present invention.
6 is a view showing the state of a crop using an image obtained through a drone flight according to an embodiment of the present invention.
7 is a flowchart illustrating a process for producing a subject map in a control center according to an embodiment of the present invention.
8 is a flowchart illustrating a process performed by the AI server in step S730 of FIG. 7 according to an embodiment of the present invention.
9 is a block diagram for explaining a learning concept for crop analysis performed in the AI server according to an embodiment of the present invention.
FIG. 10 is a conceptual diagram illustrating a method of generating an input image for image classification through a generator determined through FIG. 9 .

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. 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 published below, but may be implemented in various different forms, only these embodiments make the publication of the present invention complete, and common knowledge in the technical field to which the present invention pertains It is provided to fully inform the possessor of the scope of the invention, and the present invention is only defined by the scope of the claims. Like reference numerals refer to like elements throughout.

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 to be interpreted ideally or excessively unless clearly defined in particular.

The terminology used in this application is only used to describe specific embodiments, and is not intended to limit the present invention. The singular expression includes the plural expression unless the context clearly dictates otherwise. In the present application, terms such as “comprise” or “have” are intended to designate that a feature, number, step, operation, component, part, or combination thereof described in the specification exists, but one or more other features It should be understood that this does not preclude the existence or addition of numbers, steps, operations, components, parts, or combinations thereof.

Terms including an ordinal number such as second, first, etc. may be used to describe various elements, but the elements are not limited by the terms. The above terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the second component may be referred to as the first component, and similarly, the first component may also be referred to as the second component. and/or includes a combination of a plurality of related listed items or any of a plurality of related listed items.

The present invention relates to a crop cultivation information service platform.

The crop cultivation information service platform 10 utilizes the unmanned aerial vehicle 20 to receive image information so that it is possible to quickly and accurately identify the cultivation area. The crop cultivation information service platform 10 enables fast and accurate observation, can bring economic effects, and increase the quantity and quality of observation. Accordingly, the income of farmers can be guaranteed, the consumer price can be stabilized, the accuracy of agricultural observation can be improved, and the predictive power of supply and demand can be strengthened.

Here, fast observation can lead to quick results for crop conversion or supply and demand preparation, accurate observation enables the advancement and modernization of inaccurate observation data, and economic observation reduces the cost of observation and economic damage caused by observation errors. has the effect of In addition, the quantity and quality of observations have the effect of providing various information such as meteorological disasters, pests, and vegetation index.

The unmanned aerial vehicle 20 can investigate crops in high altitudes, acquire images to quickly and accurately observe crop cultivation area and growth, and can observe even in areas with steep terrain and severe diurnal temperature differences in areas above 850 m above sea level. have.

According to an embodiment of the present invention, the crop cultivation information service platform 10 may identify and express the seasonal and sequential changes in the cultivation area of Chinese cabbage in high altitudes through photographing using the unmanned aerial vehicle 20 . The crop cultivation information service platform 10 improves the efficiency of observation tasks through fast and accurate processing of images acquired using drones and time series change management based on GIS (Geographic Information System), Drone image data can be effectively utilized. Here, high-altitude cabbage representing a crop is an example, and is not necessarily limited thereto.

Geographic Information System (GIS) is a kind of information system that processes and manages geographic information, that is, spatially distributed information. Here, spatially distributed information, that is, geographic information is usually presented in the form of a map, and a geographic information system means an information system that can input, store, process, analyze, manage, and output such geographic information. The GIS can classify lots and derive the area based on a farm map. Here, the Farm Map may be used when extracting a boundary line for each lot.

Therefore, the crop cultivation information service platform 10 can observe time-series changes of the cultivation area, can check the economic feasibility and utility of observing the cultivation area and shipment changes through drones, and it is possible to grasp the variability by grasping real-time images, , it is possible to effectively observe crops in the open field, and to preemptively secure basic data for growth judgment, such as identifying diseases, pests, and weather damage for sophistication.

1 is an exemplary diagram illustrating a crop cultivation information service using a crop cultivation information service platform according to an embodiment of the present invention.

Referring to FIG. 1 , the crop cultivation information service platform 10 may exchange information with the unmanned aerial vehicle 20 , the first terminal 30 , and the second terminal 32 , as illustrated in FIG. 1 . Among various components, information may be exchanged by omitting some components or by additionally including other components.

Referring to FIG. 1 , the crop cultivation information service platform 10 may include a crop condition determination device 12 and a monitoring unit 14 . Here, the crop status determination device 12 may be implemented by being included in the crop cultivation information service platform 10 , but may also be implemented as a separate device.

The crop status determination device 12 is a device that receives the crop image obtained from the unmanned aerial vehicle 20 and determines the type of crop and the status of the crop. The crop cultivation information service platform 10 may generate a crop map by using the flight path the unmanned aerial vehicle 20 flew, the type of crop determined by the crop state determination device 12 and the state of the crop.

According to an embodiment of the present invention, the crop cultivation information service platform 10 may utilize the image information produced by executing the flight through the unmanned aerial vehicle 20 . The unmanned aerial vehicle 20 may acquire an image, a thermal image, a hyperspectral image, and the like and transmit it to the crop cultivation information service platform 10 . The crop cultivation information service platform 10 creates and demonstrates the flight path of the unmanned aerial vehicle 20, collects the information produced by the unmanned aerial vehicle 20, and processes big data to derive AI crop classification, cultivation, and harvest status , visualization, GIS-based analysis expression, etc., and using spatial information, the researcher can manage flight routes, collection method development (DSM 3D terrain input, Farm Map input), survey area, transmission data transmission and reception, etc. In addition, the crop cultivation information service platform 10 provides the path of the unmanned aerial vehicle 20, collection, storage, analysis, etc., to an application, public institution, research institution, company, academia, other industries, etc. that manage the unmanned aerial vehicle 20 may be provided, but is not necessarily limited thereto.

The unmanned aerial vehicle 20 performs a flight along a flight path and acquires environmental information or crop images. Here, the flight path is a path through which the unmanned aerial vehicle 20 performs flight and flies to obtain environmental information or crop images, and a flight path may be generated in consideration of a take-off point and a landing point.

The first terminal 30 and the second terminal 32 are electronic devices that transmit data required for processing by the crop cultivation information service platform 10 and the unmanned aerial vehicle 20 and receive the processed data. The first terminal 30 and the second terminal 32 may be implemented as a computing device, and a smart phone, a personal computer, a tablet PC, a portable personal information terminal ( Personal Digital Assistant (PDA), laptop (Laptop), etc. may be there, but is not limited thereto.

The crop cultivation information service platform 10 , the unmanned aerial vehicle 20 , the first terminal 30 , and the second terminal 32 are connected through a communication network. The communication network means a set of communication facilities connected for the purpose of enabling communication between the crop cultivation information service platform 10 , the unmanned aerial vehicle 20 , the first terminal 30 , and the second terminal 32 . The communication network includes nodes, lines, trunk lines, and satellites, and these are connected and connected to each other.

The crop cultivation information service platform 10 , the unmanned aerial vehicle 20 , the first terminal 30 and the second terminal 32 are capable of wired and wireless communication. For example, various communication protocols such as short-range wireless communication, long-distance wireless communication, mobile communication, and wireless LAN communication may be used for wireless communication. Examples of wireless communication protocols include Near Field Communication (NFC), ZigBee, Bluetooth, Wi-Fi, WiMAX, Global System For Mobile Communication (GSM), 3G (Third Generation) There are mobile communication, Long Term Evolution (LTE), 4G, 5G, etc., but is not limited thereto.

The crop cultivation information service platform 10 includes a database. The database refers to a data storage form in which data search, extraction, deletion, editing, addition, and the like can be freely performed. Databases are Oracle, Infomix, Sybase, Relational Data Base Management System (RDBMS), Gemston, Orion, object-oriented database management system ( Object Oriented Database Management System (OODBMS) may be implemented according to the purpose of the present embodiment by using a distributed database, cloud, or the like.

The crop cultivation information service platform 10 includes a processor and a memory for storing a program executed by the processor. Here, the crop cultivation information service platform 10 may be implemented as a crop cultivation information service server.

The crop cultivation information service platform 10 may perform processes to be described later by a processor, and data according to the process may be stored in a memory, but is not limited thereto.

According to an embodiment of the present invention, the processor is a crop cultivation information service platform 10, the unmanned aerial vehicle 20, the first terminal 30 and the second terminal 32 behavior history by the act of exchanging information with each other can be created and saved. For example, the behavior history may include a crop map provided by the crop cultivation information service platform 10 to the first terminal 30 and the second terminal 32, a flight path provided to the unmanned aerial vehicle 20, etc. Through this, it is possible to check the crop map provision status for each region, the history of the flight of the unmanned aerial vehicle 20 for a certain period, and the like.

According to an embodiment of the present invention, the unmanned aerial vehicle 20 and the crop cultivation information service platform 10 exchange data in real time, and use 4G, 5G wireless networks and Wifi to instantly transmit images, and delete data. A transmission function can be performed.

The unmanned aerial vehicle 20 may transmit the acquired crop image to the crop cultivation information service platform 10 in real time, and may additionally transmit latitude, longitude, state information, gimbal state information, and the like.

The first terminal 30 is a terminal used by a user to check the unmanned aerial vehicle 20, and can check image information and flight information in real time through a real-time app, and the unmanned aerial vehicle 20 and OcuSync, Lightbridge2, WiFi, etc. communication can be performed through

The second terminal 32 is a terminal capable of monitoring a crop map that visualizes cultivation area, harvest status, growth, pests, and disaster information. It may be a terminal used, etc., but is not necessarily limited thereto.

The crop cultivation information service platform 10 may generate a flight path of the unmanned aerial vehicle 20 according to the land formed in consideration of preset topographic information and take-off and landing points. The crop cultivation information service platform 10 receives and analyzes the crop image acquired by the unmanned aerial vehicle 20 to obtain crop information according to the flight path, and based on the flight path and crop information, the harvest status and disaster status by cultivation area You can create a crop map according to

The crop cultivation information service platform 10 extracts a representative point of the lot indicating the center point of the lot based on preset topographic information including the lot information on the lot, and takes the flight in consideration of the flight trajectory according to the extracted representative point. You can create a route.

The lot information may include division boundary information between parcels and altitude information of areas included in the lot.

The crop cultivation information service platform 10 may set the take-off and landing points of the unmanned aerial vehicle 20 when creating a flight path. Here, the take-off and landing point is a point where both take-off and landing are possible. The take-off and landing point can be set to a place with easy access, no telephone poles, trees, etc., and easy visibility. In addition, take-off and landing points may be set where they do not interfere with the passage of residents' vehicles.

According to an embodiment of the present invention, two take-off and landing points may be set including a take-off point and a landing point, but the present invention is not limited thereto, and at least one take-off and landing point may be set. The unmanned aerial vehicle 20 can safely perform take-off and landing through at least one take-off and landing point.

The take-off and landing points may be set in the crop cultivation information service platform 10 , and coordinates of the take-off and landing points may be transmitted to the unmanned aerial vehicle 20 .

In addition, the unmanned aerial vehicle 20 may acquire a crop image while flying a plurality of sub-regions. In this case, the sub-region may be divided and set in the crop cultivation information service platform 10 into at least two regions based on an area, an elevation difference, and the like.

The crop cultivation information service platform 10 may transmit the reset route except for the flight route to the unmanned aerial vehicle 20 .

The reset path may further include a path for re-photographing the crop image when crop information is not recognized through the crop image obtained according to the flight path. Specifically, the reset path may be generated by acquiring surrounding environment information while the unmanned aerial vehicle 20 moves along the flight path, and resetting the optimal flight path based on the acquired environment information. Here, the environmental information may include factors that prevent the unmanned aerial vehicle 20 from taking a crop image while flying, such as the number of obstacles, the size of the obstacles, and the weather.

The crop cultivation information service platform 10 observes the movement along the flight path of the unmanned aerial vehicle 20 in real time, and when the unmanned aerial vehicle 20 deviates from the flight path, a check signal of the unmanned aerial vehicle 20 is generated. It is possible to receive an abnormal signal or a normal signal generated by checking the status of the GPS and lidar sensors that check the position and movement direction of the unmanned aerial vehicle 20 by the inspection signal.

The crop cultivation information service platform 10 receives an abnormal signal when there is an abnormality in the state check of the unmanned aerial vehicle 20 to generate a return signal for the unmanned aerial vehicle 20 to return, and receives a normal signal when there is no abnormality Thus, by retransmitting the flight path to the unmanned aerial vehicle 20, the movement according to the flight path can be re-observed.

The crop cultivation information service platform 10 may analyze crop images through a convolutional neural network (CNN) and classify them by crops to obtain crop information including types of crops and state of crops.

The crop cultivation information service platform 10 performs primary crop recognition by analyzing the first crop image through a convolutional neural network, and when crop information is not obtained through primary crop recognition, the location of the unmanned aerial vehicle 20 Alternatively, secondary crop recognition is performed by analyzing the second crop image obtained by adjusting the shooting angle through a convolutional neural network, and when crop information is not obtained through secondary crop recognition, the first crop image and the second crop A tertiary crop recognition can be performed by synthesizing an image through an adversarial generative neural network (GAN) to generate a synthetic image, and analyzing the generated synthetic image through a convolutional neural network.

Tertiary crop recognition analyzes the synthesized image generated through an adversarial generative neural network (GAN) that synthesizes the first crop image or the second crop image and the label information of the first crop image or the second crop image through the generator as input. can be performed. The label information may include time information, location information, and color information at which the first crop image or the second crop image was obtained.

The crop cultivation information service platform 10 primarily uses a convolutional neural network for the crop image, and a plurality of classification models according to the type of crop and a plurality of classification models according to the state of the crop including normal, deficiency, high temperature damage, and disease. A state model can be learned, and a plurality of learned classification models are applied to analyze the characteristics extracted from the crop image to determine the crop type, and the state model can be applied to the determined crop type to determine the state of the crop. can

The crop cultivation information service platform 10 applies a plurality of state models to extract state feature values according to the state of the crop, compares the state feature value with a preset reference feature value, and includes normal, deficiency, high temperature damage, and disease. It is possible to determine the state of the crop, and by applying the time value applied to the state model, the preset reference characteristic value for each time value can be changed.

The crop cultivation information service platform 10 may generate a crop map in which the harvest status and disaster status for each cultivation area are displayed at a location along the flight path in consideration of crop information including the type of crop and the status of the crop.

The crop cultivation information service platform 10 calculates the cultivation area according to crop information among the total area according to the flight path, extracts the coordinates of the crop information, and generates a crop map using the flight path, crop information, cultivation area and coordinates. can

The process in which the above-described crop cultivation information service platform 10 determines the type of crop and the crop status may be performed through the crop status determination device 12 . Accordingly, the crop state determination device 12 learns a plurality of classification models according to the type of crop and a plurality of state models according to the state of the crop, and obtains the crop image obtained by the unmanned aerial vehicle 20 moving along the flight path. After receiving the information, it is possible to obtain crop information including the type of crop and the state of the crop by analyzing it through a convolutional neural network (CNN) and classifying it by crop.

The crop map may be generated by acquiring crop information while the unmanned aerial vehicle 20 acquires a crop image.

The unmanned aerial vehicle 20 of this embodiment may be implemented as a drone, a small helicopter, an airplane, or the like, and specifically, it may be implemented as an airplane moving mainly by remote control, without being controlled by a person on board.

According to an embodiment of the present invention, the crop cultivation information service platform 10 may manage crop data by week, month, and year through a crop map. Specifically, the crop cultivation information service platform 10 can compare the crop situation every month and every year, and can check the crop preference for each year, the change of the cultivation area, the crop change according to the disaster, etc., and predict the future situation based on this It can be provided to public institutions, citizens, research institutions, corporations, academia, and other industries.

As the crop cultivation information service platform 10 provides a crop map including the type of crop and the state of the crop, it is possible to predict the yield for each type of crop to be distributed in the market, and through this, it is possible to predict the market price. Specifically, the crop cultivation information service platform 10 may predict the yield for each type of crop, and predict the amount to be discarded according to the state of the crop in the yield, thereby predicting the proportion of crops that can be distributed among the total yield.

In addition, the crop cultivation information service platform 10 may set and provide an amount to be grown in order to harvest a target yield for each crop by using a distribution ratio of crops. The crop cultivation information service platform 10 provides the types of crops that need additional cultivation or the types of crops that are no longer required for cultivation based on the ratio of crops that can be distributed, manages the amount of crops grown and suggests types of crops, Accordingly, the market price can be managed stably.

2 is a block diagram illustrating and describing a computing environment including a computing device suitable for use in example embodiments. In the illustrated embodiment, each component may have different functions and capabilities other than those described below, and may include additional components other than those described below.

The illustrated computing environment includes a crop cultivation information service platform 10 . In an embodiment, the crop cultivation information service platform 10 may be any type of computing device that transmits and receives signals to and from other terminals.

The crop cultivation information service platform 10 includes at least one processor 110 , a computer-readable storage medium 120 , and a communication bus 160 . The processor 110 may cause the crop cultivation information service platform 10 to operate according to the above-mentioned exemplary embodiment. For example, the processor 110 may execute one or more programs stored in the computer-readable storage medium 120 . The one or more programs may include one or more computer-executable instructions, which, when executed by the processor 110 , cause the crop cultivation information service platform 10 to perform operations according to the exemplary embodiment. can be configured to perform.

Computer-readable storage medium 120 is configured to store computer-executable instructions or program code, program data, and/or other suitable form of information. The program 130 stored in the computer-readable storage medium 120 includes a set of instructions executable by the processor 110 . In one embodiment, computer-readable storage medium 120 includes memory (volatile memory, such as random access memory, non-volatile memory, or a suitable combination thereof), one or more magnetic disk storage devices, optical disk storage devices, It may be flash memory devices, other types of storage media accessed by the crop cultivation information service platform 10 and capable of storing desired information, or a suitable combination thereof.

Communication bus 160 interconnects various other components of crop cultivation information service platform 10 including processor 110 and computer readable storage medium 120 .

The crop cultivation information service platform 10 may also include one or more input/output interfaces 140 and one or more communication interfaces 150 that provide interfaces for one or more input/output devices (not shown). The input/output interface 140 and the communication interface 150 are connected to the communication bus 160 . The input/output device (not shown) may be connected to other components of the crop cultivation information service platform 10 through the input/output interface 140 . Exemplary input/output devices include input devices such as pointing devices (such as a mouse or trackpad), keyboards, touch input devices (such as touchpads or touchscreens), voice or sound input devices, various types of sensor devices and/or imaging devices; and/or output devices such as display devices, printers, speakers and/or network cards. An exemplary input/output device (not shown) may be included in the crop cultivation information service platform 10 as a component constituting the crop cultivation information service platform 10 , and is distinct from the crop cultivation information service platform 10 . It may be connected to the computing device as a separate device.

3 is a view sequentially illustrating a process of obtaining crop cultivation information according to an embodiment of the present invention.

Figure 3 (a) is a process of generating a flight path in the crop cultivation information service platform (10). At this time, the generated flight path is transmitted to the unmanned aerial vehicle 20 . For example, in an area where a wind power plant is located, a flight path may be set by designing to deviate from the path so as to escape the influence of the magnetic field. As such, the flight path is set in consideration of all factors that may interfere with the flight of the unmanned aerial vehicle 20 . The flight path extracts the plane coordinates of the destination and reflects the altitude change value to automatically create a three-dimensional flight path.

Figure 3 (b) is a process of acquiring a crop image while the unmanned aerial vehicle 20 flies. At this time, the flight moves along the flight path obtained in (a) of FIG.

The unmanned aerial vehicle 20 moves along a flight path to acquire a crop image, and transmits the acquired crop image to the crop cultivation information service platform 10 in real time.

3 (c) is a process of reading a crop from the crop cultivation information service platform 10 that has received the crop image and extracting the coordinates of the crop. In this case, the crop reading may be performed through artificial intelligence (AI), but is not limited thereto. Specifically, the crop cultivation information service platform 10 may perform crop reading through a convolutional neural network (CNN), which will be described in detail in FIG. 5 .

FIG. 3( d ) is a process of producing an orthodox map and a subject map in the crop cultivation information service platform 10 , and a crop map may be generated.

The orthographic map is a map made to the same scale by correcting the distortion of each point on the crop image caused by the remark of the ground surface. Specifically, the orthographic map is a diagram in which the geometric distortion of the crop image obtained from the unmanned aerial vehicle 20 is corrected due to the topographical undulations and converted into the appearance when all objects are viewed vertically, and includes coordinates and periods.

Thematic map is a map created for the purpose of expressing a specific topic. According to an embodiment of the present invention, the subject map may be a map in which crops are divided and displayed according to parcels, but is not necessarily limited thereto.

The crop cultivation information service platform 10 may read the type of crop while photographing through the unmanned aerial vehicle 20 , display it on the map, and derive the cultivation area. In this case, the crop cultivation information service platform 10 provides real-time cultivation area so that shooting and result generation can be completed at the same time, and it is possible to check the time-series crop cultivation change situation.

The unmanned aerial vehicle 20 may receive a field survey and a flight route in which takeoff and landing points are set from the crop cultivation information service platform 10 , and transmit data obtained by photographing the orthography and altitude.

The crop cultivation information service platform 10 may set GIS data, generate an optimal flight path of the unmanned aerial vehicle 20, transmit the flight path, and receive and manage data. In addition, the crop cultivation information service platform 10 analyzes the crop image to classify crops, calculate the cultivation area, secure the topic map GUI, AI learning data, input GIS spatial information, and manage the data server can do.

Here, the GUI refers to an environment in which a user can work through graphics when exchanging information with a computer.

Therefore, the crop cultivation information service platform 10 reflects the characteristics of the mountain topography survey and grasps the crop cultivation status and growth status through 3D topographic flight. The real-time image data sent from the unmanned aerial vehicle 20 can generate a subject status map on the day of shooting through image analysis, which is a technology that matches the goal of agricultural observation. accuracy can be ensured.

The crop cultivation information service platform 10 creates an orthographic image, categorizes crops through crop images, produces a cultivation map for each crop and each lot, separates cabbage cultivation fields, displays changes in cabbage cultivation fields, and watches may deteriorate. Here, in the web-based real-time time-series value plot, crop cultivation status, shipment area, distribution map of planting season, distribution map of harvest time, etc. can all be expressed in figures and values.

The crop cultivation information service platform 10 may generate an orthographic image based on the crop image obtained from the unmanned aerial vehicle 20, and the classification of crops through the crop image can classify the crops through the AI server of FIG. have.

The crop cultivation information service platform 10 may display crops classified through crop images in different colors to display crops according to each lot to produce a crop map, and only Chinese cabbage can be extracted from the obtained crop map. . Step S1440 is illustrated as an embodiment and may be separated into other crop cultivation fields other than Chinese cabbage.

The crop cultivation information service platform 10 may display a change by comparing the previously identified cabbage cultivation parcels with the cabbage cultivation parcels currently confirmed through the unmanned aerial vehicle 20 . Specifically, when the number of fields for cultivation of Chinese cabbage is increased, areas different from the existing ones may be displayed in different colors to confirm the added fields.

The crop cultivation information service platform 10 may be formed by displaying the area in which the change of the cabbage cultivation land is displayed on the orthographic map. Specifically, time seriesization is the arrangement of observation results into a series according to a certain criterion.

The crop cultivation information service platform 10 may further consider the time value when determining the type of crop and the state of the crop to observe the state of the crop in more detail. Specifically, the crop cultivation information service platform 10 may have a different color of crops for each season, and may be divided into a normal state or an abnormal state depending on the color of the crop for each season. can be known more clearly.

The web-based real-time time-series value diagram is a time-series of changes in cultivated area in real time using GIS technology and data analysis image implementation technology.

Crop cultivation information service platform 10 by crop item. It is possible to calculate the cultivated area, planted area, and shipment area for each lot with images and values, and to use the database analysis tool to implement the crop classification diagram, the planting season distribution map, and the harvest time distribution map. At this time, a low-altitude photo of the site is placed on each lot, and the site photo can be opened with a click.

4 is a diagram of extracting a representative point of lots according to an embodiment of the present invention, and generating a flight trajectory according to the extracted representative point of a lot.

The crop cultivation information service platform 10 may obtain lot information about the lot from the farmland information database (GIS). The lot information includes division boundary information between parcels and altitude information of areas included in the lot. In the present embodiment, the representative point of a lot may be a central point of each lot or a feature point that well indicates the characteristics of the lot.

Referring to FIG. 4 , the central point of each lot may be displayed as a representative point of the lot to indicate a path connecting the representative points of the lot. The path connecting the representative points of FIG. 4 may be the actual driving path of the unmanned aerial vehicle 20, but is not necessarily limited thereto.

The flight zone on the left has a total flight distance of 3.83Km, an estimated flight time of 12 minutes and 23 seconds, and a total number of points can be 73. The flight area on the right has a total flight distance of 6.21Km, an estimated flight time of 17 minutes and 12 seconds, and a total number of points can be 94. Here, the drones in the left and right flight areas have the same flight altitude of 20 m.

Accordingly, the driving path of the unmanned aerial vehicle 20 may be determined to minimize the sum of the connection distances connecting the representative points of the lot. The driving path may be a flight path determined in advance by the crop cultivation information service platform 10 , and the path may be changed while driving.

According to another embodiment of the present invention, the unmanned aerial vehicle 20 can transmit an image acquired through a camera to the external crop cultivation information service platform 10 in real time. The crop cultivation information service platform 10 may analyze the type and distribution of crops quickly or in real time through image processing for the acquired image.

The flight of the unmanned aerial vehicle 20 may be modified in consideration of existing crop information. For example, in the case of an onion, it is not possible to distinguish only with a vertical image, and it is necessary to shoot an image at 45 degrees from the side. Therefore, in determining the path of the unmanned aerial vehicle 20, in the case of a plot with a history of onion cultivation, by utilizing the history information of crops, the image is displayed at an oblique angle at a height lower than 20 meters, for example, in a section of 10 to 15 meters. It is desirable to obtain In this case, it is preferable that the position of the unmanned aerial vehicle 20 is not the representative point of the parcel, but the representative point of the parcel that the camera of the unmanned aerial vehicle 20 looks at.

5 is a diagram illustrating a concept of performing image learning through AI and recognizing a type of crop through this according to an embodiment of the present invention.

The artificial neural network can perform distinct operations of the learning phase and the application phase. In the learning step, the artificial neural network receives training data, generates a feature value according to a predetermined weight value, and classifies the type of crop using the generated feature value.

The artificial neural network may be implemented in the form of software executed in a processor. It is preferable that the training data be provided with various images for each season so as to cover various crops. The training data may already include information about crops (type of crops, season information, crop stage information, etc.), and supervised learning using this information is possible.

Through learning through training data, a weight value included in the artificial neural network may be determined within a range that minimizes errors in crop recognition. The artificial neural network may be, for example, a convolutional neural network, which is advantageous for image processing and object recognition.

The crop cultivation information service platform 10 can classify crops from seedling time through flight, and the cultivation crop confirmation process can be faster and more accurate. In addition, the collected growth data can be accumulated as AI learning data such as crop reading, growth, and pest damage data.

Specifically, the crop cultivation information service platform 10 may classify crops through a classification model of a convolutional neural network (CNN).

Convolutional Neural Network (CNN) is a structure that extracts features of crop images and identifies patterns of features. Specifically, the crop cultivation information service platform 10 performs feature extraction when a crop image is input. Feature extraction may refer to a process of extracting a feature through an operation such as a convolution operation, extending it to a nonlinear space through an activation function, and compressing it through pooling. Through this, classification according to the characteristics of objects in a high-dimensional non-linear space is possible.

A classification model extracts features through a convolution operation. It expands into a non-linear space through an activation function, and compresses features through a pooling operation. The classification model may include a first classification model for learning a first crop region belonging to a preset crop and a second classification model for learning a second crop region not belonging to a preset crop. Here, the crop cultivation information service platform 10 may learn the first classification model for learning the first crop region and the second classification model for learning the second crop region independently of each other.

The crop cultivation information service platform 10 may primarily classify crops according to types by analyzing crop images input through the unmanned aerial vehicle 20 through the learned classification model. Specifically, the classification model may learn images of cultivated crops such as cabbage, radishes, cucumbers, carrots, and the like, and may further learn not only the entire image but also the image when the cultivated crop is photographed from the top. Through this, when the crop image taken through the unmanned aerial vehicle 20 is input, the crop cultivation information service platform 10 applies a classification model to the crop image and extracts the characteristics of the crop by inputting the time value of the crop image as input. The type of crop can be determined.

The crop cultivation information service platform 10 may determine the state of the crop including normality, deficiency, high temperature damage, and disease by applying a plurality of state models learned according to the secondary determined crop type. Specifically, the state model can learn state images such as normality, deficiency, high temperature damage, disease, etc. for each cultivated crop, and can further learn not only the entire image but also the image when the cultivated crop is photographed from the top. Through this, when the type of crop is primarily determined, the crop cultivation information service platform 10 may determine the status of the crop by inputting a time value as an input according to the determined type of crop and applying the status model.

In addition, the crop cultivation information service platform 10 may obtain an image from which a color change according to season is removed from the image by applying a seasonal filter to the crop image. Here, the seasonal filter may define a seasonal filter, for example, a filter of spring (yellow), summer (yellow green), and autumn (green), which are main colors for each season, in the case of Chinese cabbage. The crop cultivation information service platform 10 may acquire the residual image from which the color according to the season is removed in consideration of the time (season) at which the crop image is acquired. For example, if the image is acquired in summer, since yellow-green is the main color, the residual image from which the yellow-green has been removed may be obtained by applying a yellow-green filter to the crop image acquired through the unmanned aerial vehicle 20 . By applying the pre-trained CNN neural network to the filtered residual image, it is possible to check whether there are any abnormalities in the current crop through color. Specifically, the residual image with the yellow-green filter removed can confirm the white, yellow, auburn, and black spots of the cabbage, and the degree of withering of the cabbage due to the high temperature, the degree of growth of the cabbage, the presence or absence of Chinese cabbage disease, etc. . Through this, it is possible to judge the situation of cabbage and prepare for it.

As described above, when the type and state of the crop is determined through the classification model and the state model, the crop cultivation information service platform 10 may generate a crop map. The crop map is generated based on the flight path of the unmanned aerial vehicle 20 , and the crop map can be generated by displaying the state of the crop according to the type of crop for each cultivation area.

6 is a view showing the state of a crop using an image obtained through a drone flight according to an embodiment of the present invention.

Fig. 6 (a) is a view showing the progress of harvest, Fig. 5 (b) is a view showing cabbage rot, Fig. 5 (c) is a view showing normal growth, Fig. 5 (d) is It is a view showing a lack of lime, and Figure 5 (e) is a view showing damage to high temperature.

The crop cultivation information service platform 10 may diagnose the progress of harvesting and the situation of various pests through the crop image acquired through the flight of the unmanned aerial vehicle 20 .

Crop cultivation information may include a type of crop, a state of a crop, and the like. The state of a crop includes the characteristics, size, and color of the crop, and through image analysis, it is possible to determine the cultivation stage of the crop and the possibility of pests. In addition, information on the harvest progress of crops can be additionally confirmed through image analysis.

According to an embodiment of the present invention, the state of the crop according to the Chinese cabbage can be determined to be in poor condition through cabbage rot, lime deficiency, high temperature damage, etc., and can be prevented by analyzing the cause thereof. Specifically, the crop cultivation information service platform 10 may determine the type of crop and the state of the crop through the crop image acquired through the unmanned aerial vehicle 20 .

The accumulated growth data can be accumulated as AI learning data to read crops, growth, and damage by pests.

According to another embodiment of the present invention, the crop cultivation information service platform 10 performs spectral imaging to obtain the vegetation index, and harvests ¼ of the biological weight at ½ of the growth period, and the measurement items after harvest are plant height, It is possible to perform a vegetation index and correlation analysis with the growth data by using the live weight, measuring the plant height and live weight of the final crop after growth, and comparing the NDVI obtained by photography with the actual crop growth data.

The vegetation index is based on the phenomenon that the spectral reflection characteristics of chlorophyll, which are directly proportional to the nitrogen content of plants, mainly absorb red and blue wavelengths while reflecting NIR (Near InfraRed) wavelengths. It is an index to evaluate the relative distribution and activity of green plants by combining Normalized Difference Vegetation Index (NDVI), which is calculated using reflectance of red and near-infrared wavelengths, can be calculated. When shooting RED-edge and Sequoir, blur caused by vibration and speed may occur, so make a work plan so that the shooting can proceed at a time when the illumination is secured. can do. It may be desirable to shoot on a bright, overcast day to avoid reflectance distortion due to direct light.

7 is a flowchart illustrating a process for producing a subject map in a control center according to an embodiment of the present invention. The process for producing the theme map is performed by the crop cultivation information service platform, and detailed descriptions and overlapping descriptions of the operations performed by the crop cultivation information service platform will be omitted.

The subject map production in the crop cultivation information service platform 10 may be performed by receiving the low-altitude crop image obtained from the low-altitude unmanned aerial vehicle (S710).

The low-altitude crop image may be transmitted to the data server via the NAS server (S720) (S722). Here, the NAS server represents a file-level computer storage device connected to a computer network.

The low-altitude crop image may be transmitted to the AI server to perform crop reading (S730). In addition, the low-altitude crop image may be further transmitted to an image reading source to perform crop reading (S732).

Step S730 may crawl image sources such as collecting photos of crops observed on public homepages such as the Rural Economic Research Institute or the Rural Development Administration, acquiring drone images of agricultural land, and due diligence when cultivating cleared surfaces. Crawling is a technology that collects documents distributed and stored in countless computers and includes them as an index of a search target. What type of technology is included in the search target and how quickly is a factor that determines the superiority.

In step S730, the crop may be read through a reading algorithm by crawling the image source. The reading algorithm may use Vggnet, Alexnet, Squeeznet, Resnet, Densenet, and the like.

In this case, the accuracy may be improved through cross-validation of the crops read according to the crop readings performed in steps S730 and S732.

The crops read in steps S730 and S732 may classify the same images through image classification (S740).

In addition, in step S750, the low-altitude crop image may omit the orthogonal correction process.

When the image classification (S740) and orthographic correction (S750) are completed, a subject map can be produced (S760). The produced subject map is stored in the data server (S770).

Although it is interposed as sequentially executing each process in FIG. 7, this is merely illustrative, and those skilled in the art change the order described in FIG. 7 within the range that does not depart from the essential characteristics of the embodiment of the present invention. Alternatively, various modifications and variations may be applied by executing one or more processes in parallel or adding other processes.

8 is a flowchart illustrating a process performed by the AI server in step S730 of FIG. 7 according to an embodiment of the present invention. The process performed by the AI server is performed by the crop cultivation information service platform, and detailed descriptions and overlapping descriptions of the crop cultivation information service platform will be omitted.

The AI server may acquire an image (S810). Here, the image is a low-altitude crop image obtained from a low-altitude unmanned aerial vehicle.

When the AI server acquires the image, the AI server may extract the image (S820). In the image extraction of step S820, as shown in the drawing of FIG. 8 , the crop image is extracted as much as a preset area from the low-altitude crop image obtained from each lot. For example, image extraction is an image located in the middle region of the acquired image, and may be extracted as a rectangular region.

In step S830, the AI server can classify crops through artificial intelligence by extracting crop images from low-altitude crop images obtained from each lot. The classification of crops through artificial intelligence can be confirmed through FIG. 5 , and each crop can be classified according to type.

In step S830, crop classification is not performed, and the process returns to step S810 of obtaining an image, and when crop classification is completed, coordinates of the image may be extracted in step S840.

In step S840, image coordinate extraction may extract coordinates according to each lot through GIS through the obtained low-altitude crop image, or low-altitude crop image shooting in each lot through GPS attached to the low-altitude unmanned aerial vehicle Coordinates can be extracted through the location of the unmanned aerial vehicle in the case of

When the coordinate extraction is completed, each region may be calculated in step S850. The area is calculated to create a crop map and can be checked through GIS based on each lot.

According to an embodiment of the present invention, data for each growth stage for AI learning can check seedling, early growth, maturity, harvest, etc.

Although it is interposed as sequentially executing each process in FIG. 8, it is only illustratively described, and those skilled in the art change the order described in FIG. Alternatively, various modifications and variations may be applied by executing one or more processes in parallel or adding other processes.

Therefore, the crop cultivation information service platform 10 transmits the drone-photographed crop image, sends the drone-photographed image to a specific folder on the server, creates an artificial intelligence deep learning reading model, operates the reading model, and compares each result value. Python code that reads crops, a module that stores and manages the read values in a database, a module that displays the reading results stored in the database on a map, a module that displays a picture of the crop and the read results when a location on the map is selected, etc. may be configured to perform the above-described processes, but is not necessarily limited thereto.

9 is a block diagram for explaining a learning concept for crop analysis performed in the AI server according to an embodiment of the present invention.

A generative neural network (GAN) is a machine learning that automatically generates images close to the real thing while a generative model and a discriminant model compete.

9 shows a learning concept performed in a processor of an AI server as a type of adversarial neural network. The processes performed by the adversarial neural network shown in FIG. 9 are stored in the memory of the AI server as a series of programs for performing the adversarial neural network, and include the following operations performed by the processor of the AI server.

First, the input information 910 for learning the adversarial neural network includes a plurality of training images obtained from a corresponding lot. As additional input information 910 , time information (climate information, shooting time) when the corresponding image was acquired, location information (location information on GPS, identification information of the corresponding lot, etc.), color information (image acquired from the corresponding lot) The main color information in ) is further included as label information (which is also training data).

As shown in FIG. 9 , a generator 930 generates a fake image by receiving label information 910 and random noise 920 as inputs. The classifier 940 is composed of a plurality of convolution filters, an activation layer, and a pooling layer, and determines Real/Fake using the final feature value.

The update of the filter coefficients of the classifier 940 and the generator 930 is feedback on the discrimination result (Real/Fake response) of the fake image generated through the generator 930 and the training image 950 that is a real sample. Through this, the filter coefficients are updated in a direction to obtain a more accurate classification result. Through this learning process, filter coefficients of the generator 930 may be determined. From a physical point of view, the classifier 940 and the generator 930 substantially correspond to a processor, and may be understood as terms representing detailed operations performed by the processor.

FIG. 10 is a conceptual diagram illustrating a method of generating an input image for image classification ( S740 ) through the generator 930 determined through FIG. 9 .

If the image acquired through the current drone can be classified with a sufficient confidence level, there is little need to perform this process. However, when it is difficult to specify the type of crop only with the currently acquired image, it is preferable to determine the type of crop using a synthetic image generated through an adversarial neural network.

In FIG. 10 , a generator 930 means a generator 930 whose learning has been completed through FIG. 9 . Here, the inputs are the current acquired image 1010 and a label 910 for the corresponding acquired image 1010 . The information 910 included in the label may include time information, location information, and color information at the time of capturing the image. Inputs for image classification (S740) are a synthesized image 960 synthesized through a generator 930 with label information 910 as an input, and a currently acquired image 1010. If the feature value extracted from the synthesized image 960 and the feature value extracted from the acquired image 1010 are summed according to a predetermined weight and then image classification is performed using the summed feature value, then the image is only obtained from the acquired image. When compared with the case where classification is performed, the reliability of classification can be further increased.

Even if all the components constituting the embodiment of the present invention described above are described as being combined or operated in combination, the present invention is not necessarily limited to this embodiment. That is, within the scope of the object of the present invention, all the components may operate by selectively combining one or more.

The above description is merely illustrative of the technical idea of the present invention, and those of ordinary skill in the art to which the present invention pertains may make various modifications, changes and substitutions within the scope without departing from the essential characteristics of the present invention. will be. Accordingly, the embodiments disclosed in the present invention and the accompanying drawings are for explaining, not limiting, the technical spirit of the present invention, and the scope of the technical spirit of the present invention is not limited by these embodiments and the accompanying drawings . The protection scope of the present invention should be construed by the following claims, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present invention.

10: Crop Cultivation Information Service Platform
12: Crop condition judgment device
14: monitoring unit
20: unmanned aerial vehicle
30: first terminal
32: second terminal

Claims (15)

In the crop cultivation information service platform comprising a processor and a memory for storing a program executed by the processor,
The processor is
Creates the flight path of the unmanned aerial vehicle according to the land formed in consideration of the preset topographic information and the take-off point and landing point,
It receives and analyzes the crop image acquired by the unmanned aerial vehicle to obtain crop information according to the flight path,
In consideration of the crop information including the type of crop and the state of the crop, a crop map is generated in which the harvest status and disaster status by cultivation area according to the flight path are displayed,
The processor learns a plurality of classification models according to the type of crop and a plurality of state models according to the state of the crop, and analyzes the crop image through a convolutional neural network (CNN) to classify each crop. Obtaining the crop information including the type of crop and the state of the crop,
The processor includes a first classification model for learning a first crop region belonging to a preset crop and a second classification model for learning a second crop region not belonging to a preset crop primarily By applying a model, inputting the time value of the crop image as an input, analyzing the extracted characteristics to determine the type of crop, and secondarily, by inputting the time value according to the determined type of crop, the plurality of learned To generate the crop map by applying a state model to determine the state of the crop, including normal, deficiency, and high temperature damage, and displaying the state of the crop according to the type of crop for each cultivation area based on the flight path Crop cultivation information service platform, characterized in that. According to claim 1,
The processor is
Extracting a representative point of the lot indicating the center point of the lot based on the preset topographic information including lot information about the lot, and generating the flight path in consideration of the flight trajectory according to the extracted representative point,
The lot information is a crop cultivation information service platform, characterized in that it includes information on the boundary between parcels and altitude information of areas included in the lot. 3. The method of claim 2,
The processor is
Create a reset route of the unmanned aerial vehicle that has been reset except for the flight route,
The reset path, if the crop information is not recognized through the crop image acquired according to the flight path, crop cultivation information service platform, characterized in that it further comprises a path for re-photographing the crop image. According to claim 1,
The processor is
Observing the movement of the unmanned aerial vehicle along the flight path in real time, and generating a check signal of the unmanned aerial vehicle when the unmanned aerial vehicle deviates from the flight path,
A crop cultivation information service platform, characterized in that it receives an abnormal signal or a normal signal generated by checking the status of a GPS and lidar sensor that confirms the position and movement direction of the unmanned aerial vehicle by the inspection signal. 5. The method of claim 4,
The processor is
When there is an abnormality in the state check of the unmanned aerial vehicle, the abnormal signal is received to generate a return signal for the unmanned aerial vehicle to return, and if there is no abnormality, the normal signal is received and the flight path is retransmitted to the unmanned aerial vehicle. Crop cultivation information service platform, characterized in that re-observing the movement along the flight path. delete delete According to claim 1,
The processor is
First crop recognition is performed by analyzing the first crop image through the convolutional neural network,
When crop information is not obtained through the primary crop recognition, the second crop image obtained by adjusting the position or shooting angle of the unmanned aerial vehicle is analyzed through the convolutional neural network to perform secondary crop recognition,
When crop information is not obtained through the secondary crop recognition, a synthesized image is generated by synthesizing the first crop image and the second crop image through an adversarial generating neural network (GAN), and the generated synthesized image is used in the A crop cultivation information service platform, characterized in that it performs tertiary crop recognition by analyzing it through a convolutional neural network. 9. The method of claim 8,
The tertiary crop recognition is
Analyze the synthesized image generated through an adversarial generative neural network (GAN) that synthesizes the first crop image or the second crop image and the label information of the first crop image or the second crop image as inputs through a generator is performed by
The label information is a crop cultivation information service platform, characterized in that it includes time information, location information, and color information at which the first crop image or the second crop image was obtained. According to claim 1,
The processor is
The unmanned aerial vehicle uses the flight path, the crop information, the cultivation area, and the coordinates to extract the crop information by calculating the cultivation area according to the crop information among the total area along the flight path. generating the crop map generated by acquiring the crop information at the same time as acquiring,
The processor is
Predicting the yield for each type of crop distributed in the market and the amount to be discarded according to the condition of the crop by using the crop map in which the type of crop and the condition of the crop are displayed,
A crop cultivation information service platform, characterized in that predicting a circulating crop ratio using the yield for each type of crop and an amount discarded according to the state of the crop, and predicting a market price using the circulating crop ratio. In the crop status determination apparatus comprising a processor and a memory for storing a program executed by the processor,
The processor is
Learning a plurality of classification models according to the type of crop and a plurality of state models according to the state of the crop,
It receives crop images acquired by an unmanned aerial vehicle moving along the flight path and analyzes it through a convolutional neural network (CNN) to classify each crop to obtain crop information including the type of crop and the state of the crop.
The processor includes a first classification model for learning a first crop region belonging to a preset crop and a second classification model for learning a second crop region not belonging to a preset crop primarily By applying a model, inputting the time value of the crop image as an input, analyzing the extracted characteristics to determine the type of crop, and secondarily, by inputting the time value according to the determined type of crop, the plurality of learned Crop condition determination device, characterized in that by applying the condition model to determine the condition of the crop including normal, deficiency, high temperature damage, and disease. delete 12. The method of claim 11,
The processor is
By applying the plurality of state models to extract a state feature value according to the state of the crop, and comparing the state feature value with a preset reference feature value to determine the state of the crop including normal, deficiency and high temperature damage, ,
Crop condition determination apparatus, characterized in that by applying the time value, the preset reference characteristic value is changed for each time value. 12. The method of claim 11,
The processor is
First crop recognition is performed by analyzing the first crop image through the convolutional neural network,
When crop information is not obtained through the primary crop recognition, the second crop image obtained by adjusting the position or shooting angle of the unmanned aerial vehicle is analyzed through the convolutional neural network to perform secondary crop recognition,
When crop information is not obtained through the secondary crop recognition, a synthesized image is generated by synthesizing the first crop image and the second crop image through an adversarial generating neural network (GAN), and the generated synthesized image is used in the Crop condition determination device, characterized in that the tertiary crop recognition is performed by analysis through a convolutional neural network. 15. The method of claim 14,
The tertiary crop recognition is
Analyzes the synthesized image generated through an adversarial generative neural network (GAN) that synthesizes the first crop image or the second crop image and the label information of the first crop image or the second crop image as inputs through a generator is performed by
The label information is a crop condition determination device, characterized in that it includes time information, location information, and color information at which the first crop image or the second crop image was obtained.

Images