KR20220045695A - Leaf wetness duration estimation method using satellite imagery and artificial intelligence - Google Patents

Leaf wetness duration estimation method using satellite imagery and artificial intelligence Download PDF

Info

Publication number
KR20220045695A
KR20220045695A KR1020200128624A KR20200128624A KR20220045695A KR 20220045695 A KR20220045695 A KR 20220045695A KR 1020200128624 A KR1020200128624 A KR 1020200128624A KR 20200128624 A KR20200128624 A KR 20200128624A KR 20220045695 A KR20220045695 A KR 20220045695A
Authority
KR
South Korea
Prior art keywords
satellite image
artificial intelligence
data
image data
wetting period
Prior art date
Application number
KR1020200128624A
Other languages
Korean (ko)
Other versions
KR102487666B1 (en
Inventor
신주영
박준상
김규랑
김부요
Original Assignee
대한민국(기상청 국립기상과학원장)
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 대한민국(기상청 국립기상과학원장) filed Critical 대한민국(기상청 국립기상과학원장)
Priority to KR1020200128624A priority Critical patent/KR102487666B1/en
Publication of KR20220045695A publication Critical patent/KR20220045695A/en
Application granted granted Critical
Publication of KR102487666B1 publication Critical patent/KR102487666B1/en

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01W—METEOROLOGY
    • G01W1/00—Meteorology
    • G01W1/10—Devices for predicting weather conditions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/22—Indexing; Data structures therefor; Storage structures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01W—METEOROLOGY
    • G01W1/00—Meteorology
    • G01W2001/006—Main server receiving weather information from several sub-stations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10032—Satellite or aerial image; Remote sensing

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Software Systems (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Environmental & Geological Engineering (AREA)
  • Ecology (AREA)
  • Computational Linguistics (AREA)
  • Biodiversity & Conservation Biology (AREA)
  • Databases & Information Systems (AREA)
  • Environmental Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Atmospheric Sciences (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Image Processing (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The present invention relates to a method for predicting a foliar wetting period using satellite image data and artificial intelligence techniques. According to the present invention, by applying the satellite image data and the artificial intelligence techniques, information on the foliar wetting period can be obtained at locations where no foliar wetting period observation equipment is installed. In addition, for a problem of inhomogeneous quality of observation data of the foliar wetting period caused by no standardization method, a single model can be applied to a wide area, so that information on the foliar wetting period of homogeneous quality can be secured. The method for predicting the foliar wetting period using the satellite image data and the artificial intelligence techniques of the present invention comprises: a first construction step (S10); a preparation step (S20); a second construction step (S30); and a prediction step (S40).

Description

위성영상자료와 인공지능기법을 이용한 엽면습윤기간 예측 방법{LEAF WETNESS DURATION ESTIMATION METHOD USING SATELLITE IMAGERY AND ARTIFICIAL INTELLIGENCE}Foliar wetting period prediction method using satellite image data and artificial intelligence technique

본 발명은 위성영상자료와 인공지능기법을 이용한 엽면습윤기간 예측 방법에 관한 것으로, 더욱 상세하게는 천리안 정지궤도의 실시간 위성자료와 인공지능 기법을 이용하여 지상면에 존재하는 식물의 엽면에 습기가 존재하는 시간을 예측하는 위성영상자료와 인공지능기법을 이용한 엽면습윤기간 예측 방법에 관한 것이다.The present invention relates to a method for predicting the foliar wetting period using satellite image data and artificial intelligence techniques. It relates to a method of predicting the foliar wetting period using satellite image data to predict the time of existence and artificial intelligence techniques.

식물의 전염병은 농업 및 임업 생산량 증대에 악영향을 주는 주요 원인이 된다.Infectious diseases of plants are a major cause of adverse effects on agricultural and forestry production growth.

박테리아와 진균류에 의한 식물 전염병은 엽면습윤기간[식물의 잎에 존재하는 수분(엽면습윤)의 지속기간]과 특정 기온에 따라 발병률이 변하는 것으로 알려져 있다.It is known that the incidence of plant diseases caused by bacteria and fungi varies depending on the foliar wetting period [the duration of moisture (foliar wetting) in the leaves of plants] and the specific temperature.

농업 및 임업 생산량 증대를 위해서는 식물 전염병에 대한 적절한 예측기술이 개발될 필요가 있다.In order to increase agricultural and forestry production, it is necessary to develop an appropriate predictive technology for plant epidemics.

식물 전염병에 큰 영향 요인인 기온은 세계기상기구에서 정해진 규칙에 따라서 다양한 장소에서 관측되고 있어, 기온자료에 대한 접근성이 높다. 또한, 높은 접근성으로 인하여 많은 연구가 진행되어 다른 기상요소들 보다 높은 예측 정확도를 가진다.Temperature, which is a major factor influencing plant epidemics, is observed in various places according to the rules set by the World Meteorological Organization, so access to temperature data is high. In addition, due to the high accessibility, many studies have been conducted, and the prediction accuracy is higher than that of other meteorological factors.

하지만, 엽면습윤기간은 세계기상기구에서 지정한 정규 기상요소가 아니라 관측기기, 장비, 설치 위치와 같은 규칙이 정해져 있지 않다. 이런 이유로 많은 기관에서 다른 관측 방법을 사용하고 있어 자료들간의 동질성을 확보하기가 어렵다. 또한, 정규 기상요소가 아니기 때문에 대부분의 기상관측소에서 엽면습윤기간을 관측하고 있지 않다. 이런 이유로 엽면습윤기간 자료에 대한 접근성과 품질이 낮아져 식물의 전염병 연구 및 예측에 많은 문제점으로 여겨지고 있다.However, the foliar wetting period is not a regular meteorological element designated by the World Meteorological Organization, and rules such as observation equipment, equipment, and installation location are not established. For this reason, it is difficult to secure the homogeneity between the data because many institutions use different observation methods. In addition, since it is not a regular meteorological component, most meteorological stations do not observe the foliar wetting period. For this reason, the accessibility and quality of foliar wetting period data is low, which is considered a problem in the study and prediction of infectious diseases of plants.

이와 같이 위에서 언급된 문제점과 한계들을 극복하기 위하여 관측기기가 없는 위치에서도 동질한 품질을 갖는 엽면습윤기간 정보를 획득할 수 있는 방법의 개발이 요구되고 있는 현실이다.As such, in order to overcome the above-mentioned problems and limitations, the development of a method for obtaining information on the foliar wetting period having the same quality even at a location without an observation device is required.

대한민국 등록특허공보 제1114513호Republic of Korea Patent Publication No. 1114513

본 발명의 목적은 정지궤도 위성영상을 이용한 옆면습윤기간을 예측하여 엽면습윤기간 관측장비가 설치되어 있지 않는 지점에서도 엽면습윤기간 정보를 얻을 수 있는 위성영상자료와 인공지능기법을 이용한 엽면습윤기간 예측 방법을 제공함에 있다.It is an object of the present invention to predict the lateral wetting period using geostationary orbit satellite images and predict the foliar wetting period using artificial intelligence and satellite image data that can obtain foliar wetting period information even at a point where no foliar wetting period observation equipment is installed. to provide a method.

본 발명의 목적은 위성영상과 단일화된 방법으로 넓은 지역의 동질한 품질의 엽면습윤기간 정보를 얻을 수 있는 위성영상자료와 인공지능기법을 이용한 엽면습윤기간 예측 방법을 제공함에 있다.It is an object of the present invention to provide a method for predicting foliar wetting period using satellite image data and artificial intelligence techniques that can obtain information on foliar wetting period of the same quality in a wide area by a method unified with satellite imagery.

상기 목적을 달성하기 위한 본 발명에 따른 위성영상자료와 인공지능기법을 이용한 엽면습윤기간 예측 방법은,The method for predicting the foliar wetting period using satellite image data and artificial intelligence technique according to the present invention for achieving the above object,

위성영상자료 및 엽면습윤기간 관측자료를 수집하여 DB(Database)를 구축하는 제1단계와,The first step of building a DB (Database) by collecting satellite image data and observation data of the foliar wetting period;

상기 DB를 전처리하여 학습, 검증, 평가 자료를 각각 준비하는 제2단계와,A second step of pre-processing the DB to prepare learning, verification, and evaluation data, respectively;

상기 학습, 검증, 평가 자료와 인공지능모형을 이용하여 엽면습윤기간 예측모형을 구축하는 제3단계와,a third step of constructing a foliar wetting period prediction model using the learning, verification and evaluation data and the artificial intelligence model;

상기 엽면습윤기간 예측모형과 상기 위성영상자료를 이용하여 대상지역의 엽면습윤기간을 예측하는 제4단계를 포함하는 것을 그 기술적 방법상의 기본 특징으로 한다.The basic feature of the technical method is to include a fourth step of predicting the foliar wetting period of the target area using the foliar wetting period prediction model and the satellite image data.

본 발명은 위성영상자료와 인공지능 기법을 적용하여 엽면습윤기간 관측장비가 설치되어 있는 않는 위치에서의 엽면습윤기간 정보를 확보할 수 있는 효과가 있다.The present invention has the effect of securing information on the foliar wetting period at a location where no foliar wetting period observation equipment is installed by applying satellite image data and artificial intelligence techniques.

본 발명은 관측 방법의 표준화된 방안이 없어 발생하는 엽면습윤기간 관측자료의 비균질한 품질 문제에 대하여 단일 모형을 넓은 지역에 적용함으로써 동질한 품질의 엽면습윤기간 정보를 얻을 수 있는 효과가 있다.The present invention has the effect of obtaining homogeneous quality information on foliar wetting period by applying a single model to a wide area in response to the problem of inhomogeneous quality of observation data of foliar wetting period caused by the lack of a standardized observation method.

도 1은 본 발명에 따른 위성영상자료와 인공지능기법을 이용한 엽면습윤기간 예측 방법을 나타내는 흐름도.
도 2는 본 발명에 따른 위성영상자료와 인공지능기법을 이용한 엽면습윤기간 예측 방법을 설명하기 위하여 천리안 2호기의 영상자료와 랜덤포레스트(RF)와 깊은신경망(DNN)을 이용하여 개발한 엽면습윤기간 예측모형의 2020년3월1일부터 동년3월4일까지의 예측 결과 예시도.1 is a flowchart illustrating a method for predicting a foliar wetting period using satellite image data and artificial intelligence techniques according to the present invention.
2 is a foliar wetting developed using image data of Cheollian Unit 2, random forest (RF) and deep neural network (DNN) to explain the method for predicting the foliar wetting period using satellite image data and artificial intelligence technique according to the present invention. An example diagram of the forecast results from March 1, 2020 to March 4, 2020 of the period forecasting model.

본 발명에 따른 위성영상자료와 인공지능기법을 이용한 엽면습윤기간 예측 방법의 바람직한 실시예를 도면을 참조하여 설명하기로 하고, 그 실시예로는 다수 개가 존재할 수 있으며, 이러한 실시예를 통하여 본 고안의 목적, 특징 및 이점들을 더욱 잘 이해할 수 있게 된다.A preferred embodiment of the method for predicting the foliar wetting period using satellite image data and artificial intelligence technique according to the present invention will be described with reference to the drawings. to better understand the purpose, features and benefits of

도 1은 본 발명에 따른 위성영상자료와 인공지능기법을 이용한 엽면습윤기간 예측 방법을 나타내는 흐름도이다.1 is a flowchart illustrating a method for predicting a foliar wetting period using satellite image data and artificial intelligence techniques according to the present invention.

본 발명에 따른 위성영상자료와 인공지능기법을 이용한 엽면습윤기간 예측 방법은 도 1에 도시된 바와 같이 위성영상자료 및 엽면습윤기간 관측자료를 수집하여 DB(Database)를 구축하는 제1단계(S10)와, 상기 DB를 전처리하여 학습, 검증, 평가 자료를 각각 준비하는 제2단계(S20)와, 상기 학습, 검증, 평가 자료와 인공지능모형을 이용하여 엽면습윤기간 예측모형을 구축하는 제3단계(S30)와, 상기 엽면습윤기간 예측모형과 상기 위성영상자료를 이용하여 대상지역의 엽면습윤기간을 예측하는 제4단계(S40)를 포함한다.The method for predicting the foliar wetting period using satellite image data and artificial intelligence according to the present invention is the first step (S10) of collecting satellite image data and observation data of foliar wetting period as shown in FIG. ), a second step (S20) of preparing learning, verification, and evaluation data by preprocessing the DB, and a third step of constructing a foliar wetting period prediction model using the learning, verification, and evaluation data and an artificial intelligence model It includes a step (S30) and a fourth step (S40) of predicting the wetting period of a target area using the foliar wetting period prediction model and the satellite image data.

상기 제1단계(S10)는 해당 분석기간에 대한 위성영상자료를 수집하여 DB를 구축하는 단계(S11)와, 엽면습윤기간 관측장비가 설치되어 있는 지점에서 관측자료 및 관측장비의 정보를 수집하여 DB를 구축하는 단계(S12)를 포함한다.The first step (S10) includes the steps of collecting satellite image data for the analysis period and building a DB (S11), and collecting observation data and information on the observation equipment at the point where the observation equipment for the foliar wetting period is installed. It includes a step (S12) of building the DB.

구체적으로, 상기 제1단계(S10)는 해당 분석기간에 대하여 위성영상자료를 수집하여 DB를 구축하는 단계(S11)로 대상지역의 정지궤도 위성 또는 공간 및 시간에 고해상도를 가지는 관측자료를 수집한다. Specifically, the first step (S10) is a step (S11) of collecting satellite image data for the corresponding analysis period and building a DB, which collects observation data with high resolution in geostationary orbit satellites or space and time of the target area. .

위성영상자료의 시간 해상도는 1시간 이하일 때만 엽면습윤기간 측정에 사용될 수 있고, 위성관측자료는 위성에 탑재되어 있는 분광카메라를 통하여 얻어진 위성영상자료이다.When the temporal resolution of the satellite image data is less than 1 hour, it can be used to measure the foliar wetting period, and the satellite observation data are satellite image data obtained through a spectroscopic camera mounted on the satellite.

이 위성영상자료는 수백 ㎚에서 수천 ㎚ 영역의 여러 협대역밴드로 구성된 채널별 영상자료로 물체에서 반사 또는 방출시키는 복사휘도(radiance)를 관측한다. 보통 4㎛ 파장을 기준으로 그 미만은 단파채널 그 이상은 장파채널로 구분하며, 이 채널자료 모두 사용될 수 있다.This satellite image data is image data for each channel composed of several narrow bands in the range of several hundred nm to several thousand nm, and the radiance reflected or emitted from an object is observed. Generally, on the basis of 4㎛ wavelength, shorter wave channels are classified as long wave channels, and all of these channel data can be used.

그리고, 대상지역 근처에 설치되어 있는 엽면습윤기간 관측장비의 정보(위치정도, 장비의 형태 등)와 관측자료를 수집하여 DB를 구축한다.And, the DB is built by collecting information (positional accuracy, type of equipment, etc.) and observation data of the foliar wetting period observation equipment installed near the target area.

상기 제2단계(S20)는 DB에서 관측장비의 위치와 같은 위치의 위성영상픽셀을 추출하는 단계(S21)와, 위성영상픽셀과 엽면습윤기간 관측정보를 결합하는 단계(S22)와, 이와 같이 결합된 자료로부터 학습, 검증, 평가 자료를 생성하는 단계(S23)를 포함한다.The second step (S20) includes the steps of extracting a satellite image pixel at the same position as that of the observation equipment from the DB (S21), and combining the satellite image pixel with the observation information of the foliar wetting period (S22), as described above. It includes a step (S23) of generating learning, verification, and evaluation data from the combined data.

인공지능기법을 이용하여 엽면습윤을 예측하는 모형을 구축하기 위해서는 인공지능기법 학습에 사용될 학습자료와, 이 학습자료를 통해 만들어진 모형의 초매개변수를 최적화한 자료와, 입력변수 선택을 위한 검증자료와 최종으로 구축된 모형의 성능을 평가하기 위한 평가자료가 필요하다. In order to build a model to predict foliar wetting using artificial intelligence, learning data to be used for learning artificial intelligence, optimizing hyperparameters of the model created through this learning data, and validation data for input variable selection and evaluation data to evaluate the performance of the finally constructed model.

위 세 자료를 만들기 위해서는 구축된 DB에서 관측장비의 위치정보를 이용하여 같은 위치의 위성영상픽셀을 추출하여 입력자료 DB를 구축한다.To create the above three data, the input data DB is constructed by extracting satellite image pixels of the same location using the location information of the observation equipment from the built DB.

만들어진 입력자료 DB를 엽면습윤 관측자료 DB와 결합하여 최종 DB를 구축하고, 최종 DB에서 6(학습):2(검증):2(평가)의 비율로 학습, 검증, 평가 자료로 나누어 놓는다.The created input data DB is combined with the leaf wetting observation data DB to construct the final DB, and the final DB is divided into learning, validation, and evaluation data in a ratio of 6 (learning):2 (verification):2 (evaluation).

상기 제3단계(S30)는 학습 및 검증 자료를 이용하여 인공지능모형들의 초매개변수(hyperparameter)를 정하는 단계(S31)와, 인공지능모형들의 입력변수(위성 영상 채널자료)를 결정하는 단계(S32)와, 평가 자료를 이용하여 완성된 각 인공지능모형의 성능을 평가하는 단계(S33)와, 이와 같은 평가를 기반으로 최종 엽면습윤기간 예측모형을 결정하는 단계(S34)를 포함한다.The third step (S30) includes the steps of determining hyperparameters of artificial intelligence models using learning and verification data (S31), and determining input variables (satellite image channel data) of artificial intelligence models ( S32), evaluating the performance of each AI model completed using the evaluation data (S33), and determining a final foliar wetting period prediction model based on the evaluation (S34).

구축된 학습 및 검증 자료를 이용하여 다양한 인공지능 모형의 초매개변수를 최적화한다.Optimize hyperparameters of various artificial intelligence models using the built-up learning and verification data.

인공지능 모형의 종류로는 인공신경망(artificial neural network), 깊은신경망(deep neural network, DNN), 랜덤포레스트(random forest, RF), 그레디언트부스팅모형(gradient boosted model), 로지스틱 희귀분석(logistic regression), 서포트벡터머신(support vector machine)과 같은 분류(classification) 모형을 사용할 수 있다.Artificial intelligence models include artificial neural networks, deep neural networks, DNNs, random forests, RF, gradient boosted models, and logistic regression. , a classification model such as a support vector machine can be used.

분류기(classifier)로서 작동할 수 있는 모든 인공지능기법을 본 발명에서 이용할 수 있고, 인공지능기법에 따라 다른 초매개변수를 가지기 때문에 최적화 방법으로는 다양한 방법론이 적용될 수 있다.All artificial intelligence techniques that can act as a classifier can be used in the present invention, and since they have different hyperparameters depending on the artificial intelligence technique, various methodologies can be applied as an optimization method.

본 발명에서 인공지능 모형의 입력변수는 위성영상자료로서 분광카메라로 촬영된 다양한 밴드의 영상자료를 의미한다. 즉, 위성영상자료의 어떤 밴드 자료를 사용하여 엽면습윤기간을 예측하는 가를 검증 자료를 이용하여 선택하게 된다.In the present invention, the input variable of the artificial intelligence model is satellite image data and means image data of various bands captured by a spectroscopic camera. That is, which band of satellite image data is used to predict the foliar wetting period is selected using verification data.

엽면습윤은 일반적으로 밤과 일출시간에 존재하기 때문에 장파채널의 자료와 낮에 태양광에 민감한 단파채널이 엽면습윤을 표현하는 데 유리하다.Since foliar wetting generally exists at night and at sunrise, data from long-wave channels and short-wave channels sensitive to sunlight during the day are advantageous for expressing foliar wetting.

위성에 탑재된 분광카메라의 특성은 채널마다 다르기 때문에 모든 채널 자료를 입력자료로서의 적합성을 검증자료로 검증하여 선택한다.Since the characteristics of the spectroscopic camera mounted on the satellite are different for each channel, all channel data are selected by verifying their suitability as input data.

최적화된 매개변수와 입력변수 조합을 가지는 인공기법들의 엽면습윤의 예측 성능의 평가자료를 이용하여 평가한다.It is evaluated using the evaluation data of the predictive performance of foliar wetting of artificial techniques with optimized parameter and input variable combinations.

즉, 완성된 모든 인공지능 모형의 엽면습윤 예측 성능과 엽면습윤기간 예측 성능을 제곱근오차, 상대오차, 상관계수, 정확도 등과 같은 평가지표를 이용하여 평가한다.That is, the foliar wetting prediction performance and foliar wetting period prediction performance of all completed AI models are evaluated using evaluation indicators such as square root error, relative error, correlation coefficient, and accuracy.

평가결과를 기반으로 최고의 성능을 보이는 인공지능 모형을 최종 엽면습윤기간 예측 모형을 결정한다.Based on the evaluation results, the artificial intelligence model showing the best performance determines the final foliar wetting period prediction model.

상기 제4단계(S40)는 위성영상자료를 취득하는 단계(S41)와, 위성영상자료 및 인공지능모형을 이용하여 대상지역의 엽면습윤기간 정보를 제공하는 단계(S42)를 포함한다.The fourth step (S40) includes acquiring satellite image data (S41) and providing information on the foliar wetting period of the target area using the satellite image data and artificial intelligence model (S42).

즉, 실시간 위성영상자료를 취득하여 완성된 엽면습윤기간 예측 모형에 위성자료를 입력하여 위성영상이 커버하는 지역의 엽면습윤기간 정보를 출력한다.That is, real-time satellite image data is acquired and satellite data is input into the completed foliar wetting period prediction model, and foliar wetting period information of the area covered by the satellite image is output.

도 2는 본 발명에 따른 위성영상자료와 인공지능기법을 이용한 엽면습윤기간 예측 방법을 설명하기 위하여 천리안 2호기의 영상자료와 랜덤포레스트(RF)와 깊은신경망(DNN)을 이용하여 개발한 엽면습윤기간 예측모형의 2020년3월1일부터 동년3월4일까지의 예측 결과 예시도로서 위성영상자료와 인공지능 기법을 적용하여 엽면습윤기간 관측장비가 설치되어 있는 않는 위치에서의 엽면습윤기간 정보를 확보할 수 있고, 표준화 방안이 없어 발생하는 엽면습윤기간 관측자료의 비균질한 품질 문제에 대하여 단일 모형을 넓은 지역에 적용함으로써 동질한 품질의 엽면습윤기간 정보를 얻을 수 있는 모습을 보여주고 있다.2 is a foliar wetting developed using image data of Chollian Unit 2, random forest (RF) and deep neural network (DNN) to explain the method for predicting the wetting period using satellite image data and artificial intelligence techniques according to the present invention. As an example diagram of the forecast results from March 1, 2020 to March 4, 2020, of the period prediction model, information on the wetting period at a location where no foliar wetting period observation equipment is installed by applying satellite image data and artificial intelligence techniques It shows that it is possible to obtain information on the wetting period of homogeneous quality by applying a single model to a wide area in response to the problem of inhomogeneous quality of observation data of the leaf wetting period caused by the lack of a standardization method.

본 발명은 관측기기가 없는 위치에서도 동질한 품질을 갖는 엽면습윤기간을 예측할 수 있는 농업 및 임업 분야 관련의 산업에 이용 가능하다.The present invention can be used in industries related to agriculture and forestry that can predict the foliar wetting period with the same quality even at a location without an observation device.

S10 : 제1단계 - DB 구축
S20 : 제2단계 - 학습, 검증, 평가 자료 준비
S30 : 제3단계 - 엽면습윤기간 예측모형 구축
S40 : 제4단계 - 대상지역의 엽면습윤기간 예측S10: Step 1 - DB Construction
S20: Step 2 - Preparation of learning, verification, and evaluation materials
S30: Step 3 - Establishment of the foliar wetting period prediction model
S40: Step 4 - Prediction of foliar wetting period of target area

Claims (5)

위성영상자료 및 엽면습윤기간 관측자료를 수집하여 DB(Database)를 구축하는 제1단계(S10)와,
상기 DB를 전처리하여 학습, 검증, 평가 자료를 각각 준비하는 제2단계(S20)와,
상기 학습, 검증, 평가 자료와 인공지능모형을 이용하여 엽면습윤기간 예측모형을 구축하는 제3단계(S30)와,
상기 엽면습윤기간 예측모형과 상기 위성영상자료를 이용하여 대상지역의 엽면습윤기간을 예측하는 제4단계(S40)를 포함하는 것을 특징으로 하는 위성영상자료와 인공지능기법을 이용한 엽면습윤기간 예측 방법.The first step (S10) of collecting satellite image data and observation data of the foliar wetting period to build a DB (Database);
A second step (S20) of pre-processing the DB to prepare learning, verification, and evaluation data, respectively;
A third step (S30) of constructing a foliar wetting period prediction model using the learning, verification, and evaluation data and the artificial intelligence model;
Foliar wetting period prediction method using satellite image data and artificial intelligence technique, characterized in that it includes a fourth step (S40) of predicting the foliar wetting period of the target area using the foliar wetting period prediction model and the satellite image data . 제1항에 있어서,
상기 제1단계(S10)는 해당 분석기간에 대한 상기 위성영상자료를 수집하여 상기 DB를 구축하는 단계(S11)와, 엽면습윤기간 관측장비가 설치되어 있는 지점에서 관측자료 및 관측장비의 정보를 수집하여 상기 DB를 구축하는 단계(S12)를 포함하는 것을 특징으로 하는 위성영상자료와 인공지능기법을 이용한 엽면습윤기간 예측 방법.The method of claim 1,
The first step (S10) includes the steps of collecting the satellite image data for the analysis period and building the DB (S11), and the observation data and information of the observation equipment at the point where the observation equipment for the foliar wetting period is installed. Foliar wetting period prediction method using satellite image data and artificial intelligence technique, characterized in that it comprises the step (S12) of collecting and building the DB. 제2항에 있어서,
상기 제2단계(S20)는 상기 DB에서 관측장비의 위치와 같은 위치의 위성영상픽셀을 추출하는 단계(S21)와, 상기 위성영상픽셀과 엽면습윤기간 관측정보를 결합하는 단계(S22)와, 상기 결합된 자료로부터 학습, 검증, 평가 자료를 생성하는 단계(S23)를 포함하는 것을 특징으로 하는 위성영상자료와 인공지능기법을 이용한 엽면습윤기간 예측 방법.3. The method of claim 2,
The second step (S20) includes the steps of extracting a satellite image pixel at the same location as that of the observation equipment from the DB (S21), and combining the satellite image pixel with the observation information on the foliar wetting period (S22); Foliar wetting period prediction method using satellite image data and artificial intelligence technique, characterized in that it comprises the step (S23) of generating learning, verification, and evaluation data from the combined data. 제3항에 있어서,
상기 제3단계(S30)는 상기 학습 및 검증 자료를 이용하여 인공지능모형들의 초매개변수(hyperparameter)를 정하는 단계(S31)와, 상기 인공지능모형들의 입력변수(위성 영상 채널자료)를 결정하는 단계(S32)와, 상기 평가 자료를 이용하여 완성된 각 인공지능모형의 성능을 평가하는 단계(S33)와, 상기 평가를 기반으로 최종 엽면습윤기간 예측모형을 결정하는 단계(S34)를 포함하는 것을 특징으로 하는 위성영상자료와 인공지능기법을 이용한 엽면습윤기간 예측 방법.4. The method of claim 3,
The third step (S30) is a step (S31) of determining hyperparameters of artificial intelligence models using the learning and verification data, and determining an input variable (satellite image channel data) of the artificial intelligence models. Step (S32), evaluating the performance of each AI model completed using the evaluation data (S33), and determining the final foliar wetting period prediction model based on the evaluation (S34) Foliar wetting period prediction method using satellite image data and artificial intelligence technique, characterized in that. 제4항에 있어서,
상기 제4단계(S40)는 상기 위성영상자료를 취득하는 단계(S41)와, 상기 위성영상자료 및 인공지능모형을 이용하여 대상지역의 엽면습윤기간 정보를 제공하는 단계(S42)를 포함하는 것을 특징으로 하는 위성영상자료와 인공지능기법을 이용한 엽면습윤기간 예측 방법.
5. The method of claim 4,
The fourth step (S40) includes the steps of acquiring the satellite image data (S41) and providing information on the foliar wetting period of the target area using the satellite image data and the artificial intelligence model (S42). A method of predicting leaf wetting period using satellite image data and artificial intelligence.
KR1020200128624A 2020-10-06 2020-10-06 Leaf wetness duration estimation method using satellite imagery and artificial intelligence KR102487666B1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
KR1020200128624A KR102487666B1 (en) 2020-10-06 2020-10-06 Leaf wetness duration estimation method using satellite imagery and artificial intelligence

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
KR1020200128624A KR102487666B1 (en) 2020-10-06 2020-10-06 Leaf wetness duration estimation method using satellite imagery and artificial intelligence

Publications (2)

Publication Number Publication Date
KR20220045695A true KR20220045695A (en) 2022-04-13
KR102487666B1 KR102487666B1 (en) 2023-01-12

Family

ID=81214973

Family Applications (1)

Application Number Title Priority Date Filing Date
KR1020200128624A KR102487666B1 (en) 2020-10-06 2020-10-06 Leaf wetness duration estimation method using satellite imagery and artificial intelligence

Country Status (1)

Country Link
KR (1) KR102487666B1 (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20090001614A (en) * 2007-05-03 2009-01-09 대한민국(산림청 국립수목원장) Method for searching wet land position based on gis
KR101114513B1 (en) 2008-10-27 2012-02-24 서울대학교산학협력단 System and method for forecasting the burst of soilborne plant diseases
US20170364816A1 (en) * 2017-08-29 2017-12-21 aWhere, Inc. Prediction of Discernible Fungus Establishment on a Crop
US20190230875A1 (en) * 2015-08-05 2019-08-01 Clearag, Inc. Customized land surface modeling in a soil-crop system using satellite data to detect irrigation and precipitation events for decision support in precision agriculture

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20090001614A (en) * 2007-05-03 2009-01-09 대한민국(산림청 국립수목원장) Method for searching wet land position based on gis
KR101114513B1 (en) 2008-10-27 2012-02-24 서울대학교산학협력단 System and method for forecasting the burst of soilborne plant diseases
US20190230875A1 (en) * 2015-08-05 2019-08-01 Clearag, Inc. Customized land surface modeling in a soil-crop system using satellite data to detect irrigation and precipitation events for decision support in precision agriculture
US20170364816A1 (en) * 2017-08-29 2017-12-21 aWhere, Inc. Prediction of Discernible Fungus Establishment on a Crop

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
Akhil Jayant Patil 외 2명. Forecasting disease spread to reduce crop losses. University of Münster (WWU), (2018.12.31.) 1부.* *
박준상, 서윤암, 김규랑, 하종철. 제주 감귤 과수원에서의 이슬지속시간 예측 모델 평가. 한국농림기상학회지. 제20권 제3호(2018), pp.262~276. (2018.09.30.) 1부.* *

Also Published As

Publication number Publication date
KR102487666B1 (en) 2023-01-12

Similar Documents

Publication Publication Date Title
US20200256978A1 (en) Crop classification and growth tracking with synthetic aperture radar
US9311605B1 (en) Modeling of time-variant grain moisture content for determination of preferred temporal harvest windows and estimation of income loss from harvesting an overly-dry crop
US9009087B1 (en) Modeling the impact of time-varying weather conditions on unit costs of post-harvest crop drying techniques using field-level analysis and forecasts of weather conditions, facility metadata, and observations and user input of grain drying data
US9076118B1 (en) Harvest advisory modeling using field-level analysis of weather conditions, observations and user input of harvest condition states, wherein a predicted harvest condition includes an estimation of standing crop dry-down rates, and an estimation of fuel costs
US9037521B1 (en) Modeling of time-variant threshability due to interactions between a crop in a field and atmospheric and soil conditions for prediction of daily opportunity windows for harvest operations using field-level diagnosis and prediction of weather conditions and observations and user input of harvest condition states
US9201991B1 (en) Risk assessment of delayed harvest operations to achieve favorable crop moisture levels using field-level diagnosis and forecasting of weather conditions and observations and user input of harvest condition states
Ballesteros et al. Applications of georeferenced high-resolution images obtained with unmanned aerial vehicles. Part II: application to maize and onion crops of a semi-arid region in Spain
Mancipe-Castro et al. Prediction of environment variables in precision agriculture using a sparse model as data fusion strategy
US9031884B1 (en) Modeling of plant wetness and seed moisture for determination of desiccant application to effect a desired harvest window using field-level diagnosis and forecasting of weather conditions and observations and user input of harvest condition states
Stümer et al. Spatial interpolation of in situ data by self-organizing map algorithms (neural networks) for the assessment of carbon stocks in European forests
Bukowiecki et al. High-throughput prediction of whole season green area index in winter wheat with an airborne multispectral sensor
López et al. A smart farming approach in automatic detection of favorable conditions for planting and crop production in the upper basin of Cauca River
Servia et al. Operational framework to predict field level crop biomass using remote sensing and data driven models
Ghamghami et al. Comparison of data mining and GDD-based models in discrimination of maize phenology
Monsiváis-Huertero et al. Assessment of NASA SMAP soil moisture products for agricultural regions in Central Mexico: An analysis based on the THEXMEX dataset
Vanli et al. Area estimation and yield forecasting of wheat in southeastern turkey using a machine learning approach
Pandey et al. Precision farming and its application
KR102487666B1 (en) Leaf wetness duration estimation method using satellite imagery and artificial intelligence
Fishman et al. Digital Villages: A Data-Driven Approach to Precision Agriculture in Small Farms.
Rathod et al. An improved space-time autoregressive moving average (STARMA) model for modelling and forecasting of spatio-temporal time-series data
Deforce et al. Harnessing the power of transformers and data fusion in smart irrigation
Yang et al. Crop harvest forecast via agronomy-informed process modelling and predictive monitoring
Piccoli et al. A deep scalable neural architecture for soil properties estimation from spectral information
CN113343783A (en) Intelligent crop identification and growth prediction method and system
Sadri et al. FarmCan: a physical, statistical, and machine learning model to forecast crop water deficit for farms

Legal Events

Date Code Title Description
E902 Notification of reason for refusal
E701 Decision to grant or registration of patent right
GRNT Written decision to grant