KR102479804B1 - Method, device and program for measuring water level, volume, inflow and pollution level using an artificial intelligence model that reflects regional and seasonal characteristics - Google Patents

Abstract

According to an embodiment of the present invention, a method for measuring a water level, a volume, an inflow and a pollution level through a satellite image of a ground surface is disclosed. The measuring method comprises the steps of: dividing the satellite image into a plurality of preset areas and classifying the plurality of areas into an observable area and an unobservable area based on a ratio of cloud area to the total area of the area; classifying a plurality of pixels included in the observable area into a water area and a land area, generating a water area image composed of pixels classified as the water areas among the plurality of pixels, and calculating at least one of a water level, a volume, an inflow, and a pollution level based on the water area image; obtaining a predicted value of the at least one of the water level, the volume, the inflow, and the pollution level through regression analysis of geographical information corresponding to the unobservable area; and clustering a plurality of learning satellite images of the ground surface at a location corresponding to each of a plurality of observatories to generate a plurality of clusters, and generating a measurement model and an estimation model for the water level, the volume, the inflow, and the pollution level based on one of the plurality of clusters. According to the present invention, errors due to regional differences can be reduced.

Description

Method, device and program for measuring water level, volume, inflow and pollution degree using artificial intelligence model reflecting regional and seasonal characteristics REGIONAL AND SEASONAL CHARACTERISTICS}

The present invention relates to a method, apparatus, and program for measuring water level, volume, inflow, and pollution degree using an artificial intelligence model in which regional and seasonal characteristics are reflected.

Unless otherwise indicated herein, material described in this section is not prior art to the claims in this application, and inclusion in this section is not an admission that it is prior art.

As global warming accelerates, damage due to rapid climate change is rapidly increasing, and accordingly, the importance of water resource management is emerging. In addition, as environmental issues become increasingly important, interest in water quality is also increasing.

Accordingly, development of a technology for managing water resources through measurement of water level, volume, inflow, and pollution level is progressing smoothly.

Mainly, technologies using measurement values measured through observatories are being developed, but there is a problem in that it is difficult to provide accurate measurement of areas separated from the observatories.

Korean Registered Patent No. 10-1751020 (registered on June 20, 2017)

One object of the present invention is to cluster learning datasets collected by region into a plurality of clusters, and to generate a measurement model and an estimation model corresponding to each of the plurality of clusters, artificial intelligence reflecting the characteristics of the region and season. It is to provide a method, device, and program for measuring water level, volume, inflow, and pollution using a model.

Another object of the present invention is a method for measuring water level, volume, inflow and pollution using an artificial intelligence model reflecting regional and seasonal characteristics, which calculates water level, volume, inflow and pollution based on feature vectors derived from satellite images. , to provide devices and programs.

Another object of the present invention is to classify the satellite image into an observable area and an unobservable area, calculate the water level, volume, inflow and pollution level corresponding to the observable area through the satellite image of the observable area, and calculate the unobservable area. A method for measuring water level, volume, inflow and pollution using an artificial intelligence model reflecting the characteristics of regions and seasons, which calculates the water level, volume, inflow and pollution corresponding to the unobservable area through satellite images and corresponding geographic information, To provide devices and programs.

Another object of the present invention is a method, apparatus, and program for measuring water level, volume, inflow, and pollution level using an artificial intelligence model reflecting regional and seasonal characteristics, capable of detecting excessive discharge of pollutants in a set monitoring area. is to provide

According to one embodiment of the present invention, an operating method for measuring water level, volume, inflow and pollution through satellite images of the ground surface is provided.

The operation method may further include dividing the satellite image into a plurality of predetermined areas and classifying the plurality of areas into an observable area and an unobservable area based on a ratio of a cloud area to a total area of the area; A plurality of pixels included in the observable area are classified into a water area and a land area, and a water area image composed of pixels classified as water areas among the plurality of pixels is generated. Calculating at least one of an inflow amount and a degree of contamination; Obtaining a predicted value of at least one of water level, volume, inflow, and pollution degree through regression analysis of geographic information corresponding to the unobservable area; And clustering a plurality of learning satellite images of the ground surface at a position corresponding to each of the plurality of observatories to generate a plurality of clusters, and based on any one of the plurality of clusters, water level, volume, inflow and It may include generating a measurement model and an estimation model for contamination.

In addition, the step of generating a measurement model and an estimation model for water level, volume, inflow, and pollution based on any one of the plurality of clusters classifies the plurality of satellite images for learning to correspond to each of a plurality of regions. and generating a plurality of datasets corresponding to each of the plurality of regions; removing a satellite image determined to be noise from among a plurality of satellite images included in each of the plurality of datasets; performing under sampling on a plurality of datasets based on a data amount of a dataset having a minimum data amount among the plurality of datasets; and generating the plurality of clusters by clustering the plurality of datasets.

The calculating of at least one of water level, volume, inflow, and pollution based on the water area image may include extracting a plurality of color histograms corresponding to a plurality of different colors from the water area image; and determining a plurality of second feature vectors from the plurality of color histograms, and calculating a volume amount corresponding to the observable area based on the plurality of second feature vectors, The corresponding volume amount may be calculated through a random forest regressor (RFR) algorithm.

In addition, the step of calculating at least one of water level, volume, inflow, and pollution based on the water area image includes a first water level corresponding to an observable area of the first satellite image, which is a satellite image captured at a first time. Calculating a first volume amount that is a volume amount corresponding to an observable area of the first satellite image; A second water level, which is a water level corresponding to an observable area of a second satellite image, which is a satellite image taken at a second time point different from the first time point, is calculated, and a volume amount corresponding to the observable area of the second satellite image is calculated. 2 calculating volume; and calculating an inflow amount corresponding to a time between the first time and the second time through multiple regression analysis on the first water level, the first volume, the second water level, and the second volume. can

In addition, the step of obtaining a predicted value of at least one of water level, volume, inflow, and pollution level through regression analysis of geographic information corresponding to the unobservable area may include at least one prediction value determined from geographic information corresponding to the unobservable area. Calculating a first predicted value for a water level corresponding to the unobservable region through regression analysis on one independent variable; and a water level corresponding to the observable region and a kriging analysis of the first predicted value to calculate a first residual corresponding to the first predicted value, and based on the first predicted value and the first residual, Calculating a water level corresponding to the unobservable region may be included.

According to an embodiment of the present invention, the water level, volume, inflow, and pollution are calculated through a measurement model and an estimation model matched with regions. Through this, errors that may occur due to regional variations can be reduced.

According to another embodiment of the present invention, since the water level, volume, inflow and pollution are measured through satellite images, the water level, volume, inflow and pollution in an area without an observatory can be measured and provided to the user.

According to another embodiment of the present invention, since the water level, volume, inflow, and pollution degree are estimated through an estimation model, the water level, volume, inflow amount, and pollution degree of a region where clouds are formed may be estimated and provided to the user.

According to another embodiment of the present invention, it is possible to continuously monitor the monitoring area for contamination.

1 is a schematic diagram of a water level, volume, inflow and contamination measuring system according to an embodiment.
FIG. 2 is a block diagram showing functional modules of the apparatus for measuring water level, volume, inflow and contamination degree according to FIG. 1 by way of example.
FIG. 3 is a flowchart illustrating a process of generating a measurement model by the model generating unit according to FIG. 2 .
4 is a flowchart illustrating a process of generating a measurement model by the model generating unit according to FIG. 2 .
5 is a diagram conceptually illustrating a process of classifying a water area image from a satellite image by the model generating unit according to FIG. 2 .
6 is a diagram conceptually illustrating a process of extracting a feature vector from a water area image by the model generator according to FIG. 2 .
FIG. 7 is a diagram conceptually illustrating a process of generating a measurement model based on a feature vector, a weather vector, and a geographic vector by the model generator according to FIG. 2 .
8 is a flowchart illustrating a process of generating an estimation model by the model generator according to FIG. 2 .
9 is a diagram conceptually illustrating a process of classifying an observable region and an unobservable region by the region classification unit according to FIG. 2 .
10 is a flowchart illustrating a process of measuring at least one of a water level, a volume, an inflow amount, and a degree of contamination by the measuring unit according to FIG. 2 .
11 is a diagram conceptually illustrating a process of measuring an inflow amount by the measuring unit according to FIG. 2 .
12 is a flowchart illustrating a process of estimating a water level by the estimator according to FIG. 2 .
FIG. 13 is a flowchart illustrating a process of estimating volume by the estimator according to FIG. 2 .
14 is a flowchart illustrating a process of estimating an inflow amount by the estimator according to FIG. 2 .
15 is a flowchart illustrating a process of estimating the degree of contamination by the estimator according to FIG. 2 .
FIG. 16 is a flowchart illustrating a process of determining whether an abnormal pattern of contamination occurs by the abnormal pattern determination unit according to FIG. 2 .
17 is a diagram showing the hardware configuration of the device according to FIG. 1 by way of example.

Since the present invention can make various changes and have various embodiments, specific embodiments will be illustrated in the drawings and described in detail in the detailed description. However, this is not intended to limit the present invention to specific embodiments, and should be understood to include all modifications, equivalents, and substitutes included in the spirit and scope of the present invention. Like reference numerals have been used for like elements throughout the description of each figure.

Terms such as first, second, A, and B may be used to describe various components, but the components should not be limited by the terms. These terms are only used for the purpose of distinguishing one component from another. For example, a first element may be termed a second element, and similarly, a second element may be termed a first element, without departing from the scope of the present invention. The terms and/or include any combination of a plurality of related recited items or any of a plurality of related recited items.

It is understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, but other elements may exist in the middle. It should be. On the other hand, when an element is referred to as “directly connected” or “directly connected” to another element, it should be understood that no other element exists in the middle.

Terms used in this application are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly dictates otherwise. In this application, the terms "include" or "have" are intended to designate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, but one or more other features It should be understood that the presence or addition of numbers, steps, operations, components, parts, or combinations thereof is not precluded.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and unless explicitly defined in the present application, they should not be interpreted in an ideal or excessively formal meaning. don't

The 'pollution degree' used below is the concentration of dissolved oxygen (DO), biochemical oxygen demand (BOD), and suspended solids (SS), and the unit of the pollution degree may be ppm. In addition, 'contamination' is a concentration of hydrogen ions, and the unit of contamination may be pH.

As used below, 'water level' refers to the depth of a stream. As used below, 'volume' refers to the volumetric capacity of a river.

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

1 is a schematic diagram of a system for measuring water level, volume, inflow and contamination according to an embodiment.

Referring to FIG. 1, the water level, volume, inflow and pollution measurement system includes a device 100, a user terminal 200, a satellite server 300, an observatory 400, a weather server 500, and a Geographic Information System (GIS). ) server 600.

The user terminal 200 is a terminal of a user who wants to use the water level, volume, inflow and pollution level providing service, registers user information in the device 100, and provides the water level, volume, inflow and pollution level through the device 100. Several features of the service are available.

The artificial satellite server 300 stores a satellite image captured by an artificial satellite operating away from the earth's surface. In one embodiment, the artificial satellite may be an artificial satellite operating at a predetermined interval from the surface of the earth. In one embodiment, the satellite server 300 may receive satellite images from the satellite at preset time intervals. In one embodiment, the artificial satellite image may be an image of a preset resolution or a preset size.

The satellite server 300 transmits the satellite image received from the satellite to the device 100 .

The observation station 400 measures the water level, volume, inflow and contamination of the installed location. A plurality of observation stations 400 may be installed in each of a plurality of areas. In one embodiment, the observatories 400 may be installed in different numbers for each region. Observatory 400 may be installed in a place adjacent to a river and configured to measure at least one of water level, volume, inflow, and pollution degree.

The plurality of observatories 400 transmits at least one of the measured water level, volume, inflow, and pollution level to the device 100 . In one embodiment, when there is a separate station server (not shown) that receives at least one of the water level, volume, inflow and pollution level measured by the plurality of stations 400, the device 100 includes the station server ( At least one of water level, volume, inflow, and contamination may be received from (not shown).

The weather server 500 is a server that collects weather information and is connected to sensors that measure weather information in an information communication manner. In one embodiment, the meteorological information includes rainfall, cloud-formed area, superprecipitation, atmospheric instability, atmospheric motion vector, aerosol optical thickness, yellow sand optical thickness, aerosol particle size, temperature and humidity profile, cloudiness, cloud top, cloud top temperature, It may include barometric pressure, altitude, sea surface temperature, aerosol visibility, cloud optical thickness, cloud particle effective radius, shortwave radiation, longwave radiation, etc.

The weather server 500 transmits weather information to the device 100 .

The GIS server 600 stores geographic information about a preset area. The preset area may be divided into a plurality of subregions (Grid) having a preset size, and geographic information may be stored for each of the plurality of subregions. In one embodiment, the geographic information may include the width of each region of the river, the depth of each region of the river, information about a structure installed in the river, information about a discharge facility adjacent to the river, and the like with respect to the detailed region. Geographic information is not limited to the above examples, and various pieces of information related to water level, volume, inflow, and pollution may be included in geographic information.

The GIS server 500 transmits geographic information to the device 100.

Examples of the user terminal 200, the satellite server 300, the observatory 400, the weather server 500, and the GIS server 600 include a desktop computer, a laptop computer, and a notebook computer capable of communicating. (notebook), smart phone, tablet PC, mobile phone, smart watch, smart glass, e-book reader, PMP (portable multimedia player) , portable game console, navigation device, digital camera, digital multimedia broadcasting (DMB) player, digital audio recorder, digital audio player, digital video recorder ), a digital video player, a personal digital assistant (PDA), and the like.

The apparatus 100 calculates at least one of water level, volume, inflow, and pollution based on information received from the satellite server 300, the observatory 400, the weather server 500, and the GIS server 600, and the calculated It may be at least one server capable of providing the user terminal 200 with at least one of water level, volume, inflow and pollution degree.

The apparatus 100 may provide at least one of the water level, volume, inflow and pollution level to the user terminal 200 by operating a water level, volume, inflow and pollution level providing service.

The device 100 classifies a satellite image of the ground surface corresponding to the observatory 400 among a plurality of satellite images received from the satellite server 300 as a satellite image for learning, and a satellite image for learning and a satellite image for learning corresponding to the satellite image for learning. A model for measuring any one of the water level, the volume, the inflow and the degree of contamination may be created based on any one of the level, the volume, the inflow and the degree of contamination.

In addition, the apparatus 100 classifies a satellite image of the surface corresponding to the observatory 400 among a plurality of satellite images received from the satellite server 300 as a satellite image for learning, A model for measuring any one of the water level, volume, inflow and pollution can be generated based on any one of the corresponding water level, volume, inflow and pollution, and at least one of geographic information and meteorological information corresponding to the learning satellite image. .

The apparatus 100 classifies geographic information corresponding to the position of the observatory 400 among a plurality of geographic information received from the GIS server 600 as geographic information for learning, and geographic information for learning and a water level corresponding to the geographic information for learning. , A model for estimating any one of the water level, the volume and the inflow may be generated based on any one of the volume and the inflow.

The apparatus 100 may divide each of the plurality of satellite images received from the satellite server 300 into a plurality of regions, and classify the plurality of regions into an observable region and an unobservable region. The apparatus 100 may classify an area in which the ratio of the cloud area to the total area is equal to or greater than a preset standard among the plurality of areas as an unobservable area.

The apparatus 100 receives a satellite image from the satellite server 300, and uses a model for measuring any one of water level, volume, inflow, and pollution from the observable area of the satellite image to any one of the water level, volume, inflow, and pollution. you can get one.

The device 100 receives a satellite image from the satellite server 300, and calculates a predicted value of any one of the water level, volume and inflow from the unobservable region of the satellite image through an estimation model for any one of the water level, volume and inflow. can be obtained In addition, the apparatus 100 may calculate a residual for the predicted value through a kriging analysis on any one of the predicted value of the water level and the volume of the unobservable area and the water level and the volume of the observable area. there is. The apparatus 100 may correct any one predicted value of water level, volume, and inflow of the unobservable area based on the predicted value and the residual.

The apparatus 100 may receive a satellite image from the satellite server 300 and obtain a predicted value of pollution level from an unobservable area of the satellite image through a kriging analysis on the pollution level.

The apparatus 100 calculates a distribution map of any one of the water level, volume, inflow, and pollution for the entire area where the satellite image is received, based on any one of the water level, volume, inflow, and pollution of the observable area and the unobservable area. can create

The apparatus 100 may set a monitoring area among all areas collected through satellite images and generate a time-sequential pollution level pattern based on the pollution level of the monitoring area collected during a predetermined period.

The apparatus 100 may determine that an abnormal pattern has occurred when the pollution level corresponding to the monitoring area deviates from the time-sequential pollution level pattern.

Fig. 2 is a block diagram illustrating functional modules of the device according to Fig. 1 by way of example;

Referring to FIG. 2, the apparatus 100 includes a model generation unit 101 for generating a measurement model and an estimation model, an area classification unit 102 for classifying a received satellite image into an observable area and an unobservable area, and measurement The measuring unit 103 measures any one of the water level, volume, inflow, and pollution degree of the observable area through a model, and the estimation model estimates a predicted value of any one of the water level, volume, inflow amount, and pollution degree of the unobservable area through an estimation model. The government 104, a distribution map generator 105 generating a distribution map of any one of the water level, volume, inflow and pollution based on any one of the water level, volume, inflow and pollution degree of the observable area and the unobservable area, and a monitoring area It may include an abnormal pattern determination unit 106 for detecting the occurrence of an abnormal pattern in .

In one embodiment, the model generation unit 101 may generate a plurality of clusters through clustering of a plurality of satellite images for learning, and may generate a measurement model corresponding to each of the plurality of clusters. As clustering methods, K-Means Clustering, Mean-Shift Clustering, Density-Based Spatial Clustering of Applications with Noise (DBSCAN), Expectation-Maximization (EM) Clustering using Gaussian Mixture Models (GMM), Agglomerative Hierarchical Clustering, and the like can be used. .

In an embodiment, the model generator 101 may generate a plurality of clusters through clustering of a plurality of geographic information for learning, and may generate an estimation model corresponding to each of the plurality of clusters.

In one embodiment, clustering may be performed on datasets grouped by region. For example, a plurality of satellite images for learning or a plurality of geographic information for learning may be grouped by region to generate regional datasets, and clustering may be performed on the regional datasets.

In one embodiment, clustering may be performed on datasets grouped by season. For example, a plurality of satellite images for learning or a plurality of geographic information for learning may be grouped by season to create seasonal datasets, and clustering may be performed on the seasonal datasets.

In one embodiment, clustering may be performed on datasets grouped by season and region. For example, a plurality of satellite images for learning or a plurality of geographic information for learning are grouped by season and region to create datasets grouped by season and region, and clustering is performed on the datasets grouped by season and region. It can be.

Satellite images and geographic information on the surface of the earth at a location corresponding to the observatory 400 are used as learning data.

The model generation unit 101 determines a satellite image of the ground surface at a position corresponding to the plurality of observatories 400 as a learning satellite image, and any one of the determined satellite image for learning and the water level, volume, and pollution level corresponding to the learning image. Create training data by labeling the measured value of . In addition, the model generation unit 101 may generate a measurement model for measuring any one of water level, volume, and pollution degree through machine learning on a plurality of learning data. In one embodiment, the model generating unit 101 converts at least one of the determined satellite image for training, weather information and geographic information corresponding to the training image to a measured value of any one of water level, volume, and pollution level corresponding to the training image. By labeling, training data can be created.

The model generation unit 101 determines a pair of satellite images taken at two different viewpoints on the surface of the earth at positions corresponding to the plurality of observatories 400 as a pair of satellite images for learning, and determines the pair of satellite images for learning and the determined pair of satellite images for learning. Data for training is created by labeling the measured value of the inflow corresponding to the image pair. In addition, the model generation unit 101 may generate a measurement model for measuring an inflow through machine learning for a plurality of learning data. In one embodiment, the model generation unit 101 converts at least one of the determined satellite image pair for learning, weather information and geographic information corresponding to the satellite image pair for learning, and a measured value for an inflow corresponding to the pair of satellite image for learning. Labeling can generate training data.

In addition, the model generating unit 101 determines at least one of geographic information and meteorological information of a location corresponding to the plurality of observatories 400 received as learning information, and the determined learning information and the water level and volume corresponding to the learning information. One of the measured values is labeled to generate training data. In addition, the model generating unit 101 may generate an estimation model for any one of the water level and the volume through machine learning on a plurality of learning data.

In addition, the model generating unit 101 includes a predicted value of the first water level at a first time determined from the water level estimation model, a predicted value of the first volume amount at the first time determined from the volume estimation model, and an estimation of the water level. The predicted value of the second water level at the second time determined from the model and the predicted value of the second volume at the second time determined from the estimation model for the volume amount are determined as learning information, and the determined learning information and the inflow amount corresponding to the learning information are measured. Label values to create training data. In addition, the model generation (unit 101) may generate an estimation model for the amount of inflow through machine learning on a plurality of training data. The first time and the second time mean any two different times.

In addition, the model generating unit 101 may generate an estimation model for the pollution level through a kriging analysis of the pollution level values of the received plurality of observatories 400 and corresponding positions.

In an embodiment, the model generation unit 101 may determine that a training satellite image having a ratio of an area of a cloud region to an area of the entire region of the satellite images for learning is equal to or greater than a predetermined standard as noise and may remove the satellite image for learning.

The model generation unit 101 groups a plurality of training data by region and season to generate regional and seasonal datasets, generates a plurality of clusters through clustering of the regional and seasonal datasets, and generates a plurality of clusters. A measurement model and estimation model corresponding to each may be created.

Since a plurality of measurement models and estimation models corresponding to each of the clusters are generated, regional characteristics and seasonal characteristics can be well reflected in the measurement models and estimation models. Through this, it is possible to reduce errors that may occur due to regional and seasonal deviations when calculating water levels, volumes, inflows, and pollution levels.

The number of the plurality of observatories 400 installed to measure any one of water level, volume, inflow, and pollution level is different for each region. Since the satellite image, weather information, and geographic information of the ground surface corresponding to the observatory 400 are used as learning data, the data amount of the learning data varies by region.

The model generation unit 101 classifies a plurality of satellite images for training to correspond to each of a plurality of regions, generates a plurality of datasets corresponding to each of a plurality of regions, and generates the minimum data among the plurality of datasets. Under sampling may be performed on a plurality of datasets based on the data amount of the dataset having the amount. The model generator 101 may generate a plurality of clusters by clustering a plurality of undersampled datasets. The model generating unit 101 may generate a measurement model and an estimation model corresponding to each of a plurality of clusters. Since the data amount of the datasets is uniformed through undersampling, the characteristics of the dataset in an area where the amount of data is small can be reflected in the measurement model and the estimation model.

In one embodiment, clustering of the satellite images for training may be performed based on a plurality of feature vectors determined from the satellite images for training. The plurality of feature vectors may be a plurality of feature vectors used for learning the measurement model.

FIG. 3 is a flowchart illustrating a process of generating a measurement model by the model generator 101 according to FIG. 2 .

First, the model generation unit 101 receives a plurality of satellite images for learning including a ground surface image of a position corresponding to each of the plurality of observatories 400 (S110).

In addition, the model generation unit 101 classifies a plurality of training images to correspond to each of a plurality of regions, and generates a plurality of datasets corresponding to each of a plurality of regions (S120). In one embodiment, the dataset may include weather information and geographic information corresponding to the training image in addition to the training image.

In addition, the model generation unit 101 removes a satellite image for learning that is determined to be noise among a plurality of satellite images for learning included in each of a plurality of datasets (S130).

In one embodiment, the model generating unit 101 may determine that a satellite image for learning having a ratio of a cloud area to a total area of a plurality of satellite images for learning is equal to or greater than a preset standard as noise and may remove it.

In addition, the model generation unit 101 performs undersampling on a plurality of datasets based on the data amount of the dataset having the minimum data amount among the plurality of datasets from which noise has been removed (S140).

In addition, the model generating unit 101 generates a plurality of clusters by clustering the plurality of undersampled datasets (S150).

In addition, the model generation unit 101 generates a measurement model corresponding to each of a plurality of clusters (S160).

FIG. 4 is a flowchart illustrating a process of generating a measurement model by the model generator 101 according to FIG. 2 .

First, the model generation unit 101 extracts a water area image from a satellite image for learning (S161).

5 is a diagram conceptually illustrating a process in which the model generation unit 101 according to FIG. 2 classifies a water area image from a satellite image. Referring to FIG. 5 , the model generation unit 101 may classify pixels of a satellite image for learning into a water area (Flooded) and other areas (Unflooded) through semantic segmentation. The model generating unit 101 may classify pixels classified as flooded into water region images. In one embodiment, a UNet neural network or an FCN neural network may be used as a semantic segmentation method.

Referring back to FIG. 4 , the model generating unit 101 extracts a plurality of color histograms for a plurality of different colors from the water area image (S162). Also, the model generator 101 may determine a plurality of feature vectors from a plurality of color histograms (S163).

FIG. 6 is a diagram conceptually illustrating a process in which the model generating unit 101 according to FIG. 2 extracts a feature vector from a water area image. In the illustrated embodiment, the model generating unit 101 may extract a color histogram for blue, a color histogram for red, and a color histogram for yellow from pixels constituting the water area image. The model generator 101 determines a plurality of feature vectors from at least one of a color histogram for blue, a color histogram for red, and a color histogram for yellow.

The model generating unit 101 may determine a plurality of first feature vectors from a plurality of color histograms in order to generate a water level measurement model. Also, the model generation unit 101 may determine a plurality of second feature vectors from a plurality of color histograms in order to generate a measurement model for volumetric quantity. In addition, the model generation unit 101 may determine a plurality of third feature vectors from a plurality of color histograms in order to generate a measurement model for contamination.

Referring back to FIG. 4 , the model generator 101 generates learning data by matching a plurality of feature vectors and learning satellite images with information measured at the corresponding observatory 400 (S164).

The model generating unit 101 may generate first learning data by matching a plurality of first feature vectors with a water level measured at an observatory 400 corresponding to a satellite image for learning. In addition, the model generation unit 101 may generate second learning data by matching a plurality of second feature vectors with a volume measured at an observatory 400 corresponding to a learning satellite image. In addition, the model generating unit 101 may generate third learning data by matching a plurality of third feature vectors with a degree of pollution measured at an observatory 400 corresponding to a satellite image for learning.

The model generation unit 101 includes first feature vectors and second feature vectors determined from the satellite image for training at the first time, and first feature vectors and second feature vectors determined from the satellite image for training at the second time, Between the first time and the second time, fourth learning data may be generated by matching the learning satellite image with the inflow measured at the corresponding observatory 400 .

In one embodiment, the model generator 101 converts at least one of a plurality of feature vectors, meteorological information and geographic information corresponding to the satellite image for learning, and information measured at the observation station 400 corresponding to the satellite image for learning to Learning data can be generated by matching.

The model generator 101 measures at least one of a plurality of first feature vectors, a first weather vector determined from weather information, and a first geographic vector determined from geographic information at an observatory 400 corresponding to a satellite image for training. First learning data may be generated by matching with the water level. A first weather vector may be generated based on information related to water level among information included in weather information. A first geographic vector may be generated based on information associated with a water level among information included in geographic information. A plurality of first weather vectors and first geographic vectors may be used to generate first learning data.

The model generation unit 101 measures at least one of a plurality of second feature vectors, a second weather vector determined from weather information, and a second geographic vector determined from geographic information at an observatory 400 corresponding to a satellite image for training. The second learning data may be generated by matching the volume amount. A second weather vector may be generated based on volume-related information among information included in the weather information. A second geographic vector may be generated based on volume-related information among information included in geographic information. A plurality of second weather vectors and second geographic vectors may be used to generate second learning data.

In addition, the model generation unit 101 generates a measurement model for any one of water level, volume, and pollution degree through random forest analysis on a learning data set composed of learning data (S165).

7 is a diagram conceptually illustrating a process in which the model generator 101 according to FIG. 2 generates a measurement model based on feature vectors, meteorological vectors, and geographic vectors. The model generation unit 101 may generate a measurement model for any one of water level, volume and pollution level based on the feature vector. Also, the model generation unit 101 may generate a measurement model for any one of the water level and the volume based on at least one of the characteristic vector, the meteorological vector, and the geographic vector.

The model generating unit 101 may generate a water level measurement model through random forest analysis on the first learning data set composed of the first learning data. In one embodiment, the model generator 101 trains a water level measurement model to calculate a water level from a plurality of first feature vectors through a random forest regressor (RFR) algorithm. In an embodiment, the model generator 101 measures the water level so that the water level is calculated from at least one of a plurality of connected first feature vectors, a first weather vector, and a first geographic vector through a random forest regressor (RFR) algorithm. train the model

The model generation unit 101 may generate a volumetric model through random forest analysis on the second training data set composed of the second training data. In one embodiment, the model generation unit 101 trains the volumetric model so that the volumetric volume is calculated from the plurality of second feature vectors through a random forest regressor (RFR) algorithm. In one embodiment, the model generation unit 101 measures the volumetric volume so that the volumetric volume is calculated from at least one of a plurality of connected second feature vectors, a second meteorological vector, and a second geographic vector through a random forest regressor (RFR) algorithm. train the model

The model generating unit 101 may generate a contamination degree measurement model through random forest analysis on a third training dataset composed of third learning data. In one embodiment, the model generating unit 101 learns a pollution degree measurement model to calculate a degree of contamination from a plurality of third feature vectors through a random forest regressor (RFR) algorithm.

RFR (Random Forest Regressor) is a type of ensemble learning method used in regression analysis, and outputs an average prediction value from a plurality of decision trees constructed in the learning process. In addition, the model generator 101 may evaluate the learned measurement model through a test satellite image, and adjust hyperparameters of a random forest regressor (RFR) based on the evaluation result.

8 is a flowchart illustrating a process of generating an estimation model by the model generation unit 101 according to FIG. 2 .

The model generation unit 101 may generate an estimation model corresponding to each of the plurality of generated clusters.

First, the model generation unit 101 determines geographic information of a location corresponding to a satellite image for learning included in a cluster as geographic information for learning (S210).

In addition, the model generating unit 101 selects at least one independent variable from geographic information for learning, and selects a measurement value measured at an observatory corresponding to a satellite image for learning as a dependent variable (S220).

In addition, the model generating unit 101 generates a learning data set based on at least one independent variable and a dependent variable (S230). A training data set is created based on a plurality of training images included in the cluster and corresponding to a plurality of geographic information for training.

In addition, the model generation unit 101 generates an estimation model through regression analysis on the generated learning data set (S240).

In one embodiment, multiple regression analysis may be used for regression analysis.

The generated estimation model may be defined by Equation 1 below.

In Equation 1, s ₀ means the target position to estimate the water level or volume, Z _R (s ₀ ) means the water level or volume estimated at s0, q _k (s ₀ ) is independent from s ₀ variables, β _k means regression coefficients for independent variables, and p-1 means the number of independent variables. q ₀ (s ₀ ) is 1.

In one embodiment, a water level estimation model may be generated through multiple regression analysis. Various information included in the geographic information for learning can be set as independent variables, and independent variables that have a relatively large influence on the dependent variable, water level, can be determined through regression analysis on the learning dataset.

In one embodiment, an estimation model for the volumetric amount may be generated through multiple regression analysis. Various information included in the geographic information for learning can be set as independent variables, and independent variables that have a relatively large influence on the volume, which is the dependent variable, can be determined through regression analysis on the learning dataset.

FIG. 9 is a diagram conceptually illustrating a process of classifying an observable area and an unobservable area by the area classification unit 102 according to FIG. 2 .

When the satellite image is received from the artificial satellite server 300, the region classification unit 102 divides the received satellite image into a plurality of regions, and classifies the divided regions into an observable region and an unobservable region. The region classifying unit 102 may classify a region in which a ratio of a cloud area to a total area is equal to or greater than a predetermined standard among a plurality of regions as an unobservable region.

In an embodiment, the area classification unit 102 may classify a plurality of areas into an observable area and an unobservable area based on weather information received from the weather server 500 . Weather information may include information about areas created by clouds.

In an embodiment, the region classification unit 102 may classify a plurality of regions into an observable region and an unobservable region through a region classification model pre-learned to classify cloud regions. The image classification model may be generated through supervised learning using a Convolution Neural Network (CNN) algorithm.

The area classification unit 102 transmits the classified observable area to the measuring unit 103 .

The area classification unit 102 transmits the classified unobservable area to the estimation unit 104 .

FIG. 10 is a flowchart illustrating a process of measuring at least one of water level, volume, inflow, and contamination by the measurement unit 103 according to FIG. 2 .

* First, the measuring unit 103 extracts a plurality of feature vectors from the observable region (S310).

In one embodiment, when measuring the water level, the measurer 103 extracts a plurality of first feature vectors from the observable region. In the case of measuring the volumetric quantity, the measuring unit 103 extracts a plurality of second feature vectors from the observable region. When measuring the degree of contamination, the measuring unit 103 extracts a plurality of third feature vectors from the observable region. When measuring the flow rate, the measuring unit 103 extracts a plurality of first feature vectors and a plurality of second feature vectors from the observable region.

In one embodiment, when measuring the water level, the measurement unit 102 extracts a plurality of first feature vectors from the observable area, determines the first weather vector from weather information corresponding to the observable area, and A first geographic vector may be determined from geographic information corresponding to an area. In the case of measuring volume, the measurement unit 102 extracts a plurality of second feature vectors from the observable area, determines a second weather vector from weather information corresponding to the observable area, and geographic information corresponding to the observable area. A second geographic vector can be determined from When measuring the flow rate, the measurement unit 103 determines a plurality of first feature vectors and a plurality of second feature vectors in the observable region, and determines the first weather vector and the second feature vector from weather information corresponding to the observable region. A meteorological vector may be determined, and a first geographic vector and a second geographic vector may be determined from geographic information corresponding to an observable area.

In addition, the measurement unit 103 measures at least one of the water level, volume, inflow amount, and pollution level corresponding to the observable area based on the plurality of feature vectors (S320).

In one embodiment, when measuring the water level, the measurement unit 103 inputs the first feature vector to the water level measurement model and obtains the water level corresponding to the observable region as an output value. Specifically, the water level is calculated through the RFR (Random Forest Regressor) algorithm of the water level measurement model. In another embodiment, when measuring the water level, the measurer 103 inputs at least one of a plurality of first feature vectors, a first meteorological vector, and a first geographic vector to the water level measurement model, and can observe it as an output value. Acquire the water level corresponding to the area. A plurality of first weather vectors or first geographic vectors may be input.

In one embodiment, when measuring the volumetric quantity, the measurement unit 103 inputs the second feature vector to the volumetric quantity measurement model, and obtains the volumetric quantity corresponding to the observable region as an output value. Specifically, the volume is calculated through the RFR (Random Forest Regressor) algorithm of the water level measurement model. In another embodiment, when measuring volume, the measurement unit 103 inputs at least one of a plurality of second feature vectors, a second meteorological vector, and a second geographic vector to the water level measurement model, and can observe the output value. A volume amount corresponding to the area is obtained. A plurality of second weather vectors or second geographic vectors may be input.

In one embodiment, when measuring the pollution level, the measurement unit 103 inputs the third feature vector to the pollution level measurement model, and obtains the pollution level corresponding to the observable region as an output value. Specifically, the degree of contamination is calculated through the RFR (Random Forest Regressor) algorithm of the water level measurement model.

FIG. 11 is a diagram conceptually illustrating a process in which the measuring unit 103 according to FIG. 2 measures an inflow amount.

Referring to FIG. 11 , the measurement unit 103 obtains a first water level and a first volume corresponding to an observable area at a first time through a water level measurement model and a volume measurement model, and can be observed at a second time. A second water level and a second volume corresponding to the area are obtained. The first time point and the second time point are arbitrary times that are different from each other. The measurement unit 103 may calculate an inflow amount between the first time and the second time based on the difference between the first water level and the second water level and the difference between the first volume and the second volume. In one embodiment, the inflow amount may be calculated by multiple regression analysis in which the first water level, the first volume amount, the second water level, and the second volume amount are independent variables and the inflow amount is a dependent variable. An inflow measurement model may be generated through multiple regression analysis on the fourth learning dataset composed of the fourth learning data.

12 is a flowchart illustrating a process of estimating a water level by the estimator 104 according to FIG. 2 .

First, when the unobservable area is received, the estimator 104 obtains a predicted value for the water level of the unobservable area from geographic information corresponding to the unobservable area through a water level estimation model (S410).

The estimator 104 extracts independent variables having a relatively strong correlation with the water level, determined in the process of learning the water level estimation model, from geographic information corresponding to the unobservable area. In addition, the estimator 104 inputs the extracted independent variables to the water level estimation model, and obtains a predicted value for the water level of the unobservable area output from the water level estimation model.

In addition, the estimator 104 calculates a residual corresponding to the predicted value through kriging analysis (S420).

When the residual is calculated, the estimator 104 calculates the water level of the unobservable region based on the predicted value and the residual (S430).

FIG. 13 is a flowchart illustrating a process of estimating volume by the estimator 104 according to FIG. 2 .

First, when the unobservable area is received, the estimator 104 obtains a predicted value for the volume of the unobservable area from geographic information corresponding to the unobservable area through a volume estimation model (S510).

The estimator 104 extracts independent variables having a relatively strong correlation with the volumetric volume, determined in the process of learning the volumetric volume estimation model, from the geographic information corresponding to the unobservable region. In addition, the estimator 104 inputs the extracted independent variables to the volume estimation model, and obtains a predicted value for the volume of the unobservable region output from the volume estimation model.

In addition, the estimator 104 calculates a residual corresponding to the predicted value through kriging analysis (S520).

When the residual is calculated, the estimator 104 calculates the volume of the unobservable region based on the predicted value and the residual (S530).

14 is a flowchart illustrating a process of estimating an inflow amount by the estimator 104 according to FIG. 2 .

First, when the unobservable area is received, the estimator 104 performs a first water level prediction value of the unobservable area and a first water level prediction value of the unobservable area from geographic information corresponding to the unobservable area at the first time through a water level estimation model and a volume estimation model. A volume prediction value is obtained (S610).

In addition, the estimator 104 calculates a residual corresponding to each of the first water level predicted value and the first volume predicted value through kriging analysis (S620).

In addition, the estimator 104 determines the first water level and the second water level of the unobservable area at the first time based on the first water level prediction value, the first volume prediction value, and the residual corresponding to each of the first water level prediction value and the first volume prediction value. 1 Volume is calculated (S630).

In addition, the estimator 104 obtains a second water level prediction value and a second volume prediction value of the unobservable area from geographic information corresponding to the unobservable area at the second time through the water level estimation model and the volume estimation model (S640). .

In addition, the estimator 104 calculates a residual corresponding to each of the second water level predicted value and the second volume predicted value through kriging analysis (S650).

In addition, the estimator 104 determines the second water level and the second water level of the unobservable area at the second time based on the second water level predicted value, the second volume predicted value, and the residual corresponding to each of the second water level predicted value and the second volume predicted value. 2 The volume is calculated (S660).

In addition, the estimator 104 calculates an inflow amount between the first time and the second time corresponding to the unobservable region based on the first water level, the first volume, the second water level, and the second volume (S670). In one embodiment, the inflow amount may be calculated by multiple regression analysis in which the first water level, the first volume amount, the second water level, and the second volume amount are independent variables and the inflow amount is a dependent variable.

The residual described in FIGS. 12, 13, and 14 may be defined by Equation 2 below.

In Equation 2, s ₀ means a target position to estimate a value, e (s ₀ ) means a residual at s ₀ , w _i (s ₀ ) is a kriging corresponding to each of a plurality of sample positions It means a weight (Kriging weight), and e (s _i ) means a regression residual corresponding to each of a plurality of sample positions. n means the number of sample positions.

In one embodiment, a plurality of contiguous regions adjacent to the unobservable region may be used as the plurality of sample locations.

A kriging weight corresponding to each of the plurality of sample positions may be determined based on a variogram function. The variogram function is a measure of the similarity of data at a certain distance.

The empirical variogram function may be defined by Equation 3 below.

In Equation 3, γ(h) denotes a variogram function, Z(x) denotes a water level or volume of sample positions, and h denotes a distance between sample positions. n means the number of sample location pairs. In one embodiment, a distance between sample locations may be set based on a size of a preset observable area.

A theoretical variogram function may be modeled from an empirical variogram function, and a kriging weight may be determined through covariance derived from the theoretical variogram function. As the theoretical variogram function, a spherical function, an exponential function, a Gaussian function, or the like may be used.

The water level and volume of the unobservable region described in FIGS. 12, 13, and 14 may be defined by Equation 4 below.

In Equation 4, s ₀ denotes a target location for estimating a water level or volumetric amount, Z _RK (s ₀ ) denotes a water level or volumetric amount estimated in s0, and q _k (s ₀ ) is an independent variable in s ₀ , β _k means the regression coefficient for independent variables, and p-1 means the number of independent variables. q ₀ (s ₀ ) is 1. w _i (s ₀ ) means a kriging weight corresponding to each of a plurality of sample positions, and e(s _i ) means a regression residual corresponding to each of a plurality of sample positions. . n means the number of sample positions.

15 is a flowchart illustrating a process of estimating the degree of contamination by the estimator 104 according to FIG. 2 .

The estimator 104 calculates the contamination degree corresponding to the unobservable area through kriging analysis (S710).

The degree of contamination can be defined by Equation 5 below.

는 s₀에서 추정된 오염도를 의미하며, w_i(s₀)는 복수의 샘플 위치들 각각과 대응하는 크리깅 웨이트(Kriging weight)를 의미하고, Z(s_i)는 복수의 샘플 위치들 각각에서 측정된 오염도를 의미한다. n은 샘플 위치의 개수를 의미한다.In Equation 5, s ₀ means a target position to estimate a value,

Means the degree of contamination estimated at s ₀ , w _i (s ₀ ) means a kriging weight corresponding to each of a plurality of sample positions, and Z (s _i ) is a plurality of sample positions at each of Means the measured contamination. n means the number of sample positions.

In one embodiment, a plurality of observable regions adjacent to the unobservable region may be used as a plurality of sample locations.

A kriging weight corresponding to each of the plurality of sample positions may be determined based on a variogram function. The variogram function is a measure of the similarity of data at a certain distance.

The empirical variogram function may be defined by Equation 3 above.

When the observable area and the unobservable area are classified in the area classification unit 102, the water level, volume, inflow amount, and pollution level corresponding to the observable area are calculated in the measurement unit 103, and the unobservable area in the estimation unit 104 and the corresponding water level, volume, inflow and pollution degree are calculated. Through this, it is possible to calculate the water level, volume, inflow, and contamination of the entire region of the satellite image received from the satellite server 300 .

The distribution map generation unit 105 is configured to generate a distribution map for the water level, volume, inflow, and pollution for the entire area where the satellite image is received, based on the measured value corresponding to the observable area and the estimated value corresponding to the unobservable area. can

FIG. 16 is a flowchart illustrating a process in which the abnormal pattern determining unit 106 according to FIG. 2 determines whether an abnormal pattern of contamination has occurred.

First, the abnormal pattern determining unit 106 determines a monitoring area (S410).

In addition, the abnormal pattern determining unit 106 generates a time-sequential pollution level pattern based on the pollution level measured for a predetermined period of time in the monitoring area (S420). For example, based on the pollution level for 20 years, a pollution level pattern that changes time-sequentially according to spring, summer, autumn, and winter may be generated. In one embodiment, a K-Mean Clustering method may be used to generate a time-sequential contamination degree pattern.

In addition, the abnormal pattern determination unit 106 may determine that an abnormal pattern has occurred when the pollution level corresponding to the monitoring area deviates from the time-sequential pollution degree pattern (S430). For example, if an excessively high pollution level is calculated compared to the average pollution level in March of the monitoring area, it can be determined that an abnormal pattern has occurred.

Through this, it is possible to monitor abnormal patterns (eg, discharge of wastewater exceeding the standard) generated in the monitoring area.

17 is a diagram showing the hardware configuration of the device according to FIG. 1 by way of example.

Referring to FIG. 11 , the apparatus 100 includes at least one processor 110 and a memory storing instructions instructing the at least one processor 110 to perform at least one operation. (memory).

The at least one operation may include the components 101 to 106 of the above-described device 100 or other functions or operation methods.

Here, the at least one processor 110 may mean a central processing unit (CPU), a graphics processing unit (GPU), or a dedicated processor for performing methods according to embodiments of the present invention. can Each of the memory 120 and the storage device 160 may include at least one of a volatile storage medium and a non-volatile storage medium.

For example, the memory 120 may be one of a read only memory (ROM) and a random access memory (RAM), and the storage device 160 may be a flash-memory. , a hard disk drive (HDD), a solid state drive (SSD), or various memory cards (eg, a micro SD card).

In addition, the device 100 may include a transceiver 130 that performs communication through a wireless network. In addition, the device 100 may further include an input interface device 140 , an output interface device 150 , a storage device 160 , and the like. Each component included in the device 100 may be connected by a bus 170 to communicate with each other.

For example, the device 100 may include a communicable desktop computer, a laptop computer, a notebook, a smart phone, a tablet PC, and a mobile phone. , smart watch, smart glass, e-book reader, PMP (portable multimedia player), portable game device, navigation device, digital camera, DMB (digital multimedia broadcasting) player , a digital audio recorder, a digital audio player, a digital video recorder, a digital video player, a personal digital assistant (PDA), and the like.

The methods according to the present invention may be implemented in the form of program instructions that can be executed by various computer means and recorded on a computer readable medium. Computer readable media may include program instructions, data files, data structures, etc. alone or in combination. Program instructions recorded on a computer readable medium may be specially designed and configured for the present invention or may be known and usable to those skilled in computer software.

Examples of computer readable media may include hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions may include not only machine language codes generated by a compiler but also high-level language codes that can be executed by a computer using an interpreter and the like. The hardware device described above may be configured to operate with at least one software module to perform the operations of the present invention, and vice versa.

In addition, the above-described method or device may be implemented by combining all or some of its components or functions, or may be implemented separately.

Although the above has been described with reference to preferred embodiments of the present invention, those skilled in the art will variously modify and change the present invention within the scope not departing from the spirit and scope of the present invention described in the claims below. You will understand that it can be done.

Claims (1)

As a method for measuring water level, volume, inflow and pollution through satellite images of the ground surface,
dividing the satellite image into a plurality of predetermined areas and classifying the plurality of areas into an observable area and an unobservable area based on a ratio of a cloud area to a total area of the area;
A plurality of pixels included in the observable area are classified into a water area and a land area, and a water area image composed of pixels classified as water areas among the plurality of pixels is generated. Calculating at least one of an inflow amount and a degree of contamination;
Obtaining a predicted value of at least one of water level, volume, inflow, and pollution degree through regression analysis of geographic information corresponding to the unobservable area; and
A plurality of learning satellite images of the ground surface corresponding to each of the plurality of observatories are clustered to generate a plurality of clusters, and based on any one of the plurality of clusters, water level, volume, inflow, and pollution degree Including the step of generating a measurement model and an estimation model for,
Obtaining a predicted value of at least one of water level, volume, inflow, and pollution through regression analysis of geographic information corresponding to the unobservable area,
Calculating a first predicted value for a water level corresponding to the unobservable area through regression analysis on at least one first independent variable determined from geographic information corresponding to the unobservable area;
Calculating a first residual corresponding to the first prediction value through a kriging analysis of the water level corresponding to the observable region and the first prediction value; and
Calculating a water level corresponding to the unobservable area based on the first predicted value and the first residual,
How to measure.

Images