KR20140062756A - Water quality montoring method with observation satellite - Google Patents

Abstract

The present invention relates to a method for monitoring water with observation satellite images capable of analyzing and monitoring the water as a multiple regression models based on the satellite images photographed in a multiband of an observation satellite. The present invention is provided to verify the multiple regression models and update the multiple regression models according to additional input data whenever the observation satellite images are additionally inputted, thereby improving the accuracy of the multiple regression models. The present invention is provided to obtain or update the multiple regression models after selecting a band according to correlation analysis and statistics analysis, thereby reducing the error of water analysis.

Description

{WATER QUALITY MONITORING METHOD WITH OBSERVATION SATELLITE}

The present invention analyzes and monitors the quality of aquatic water using a multiple regression model based on a satellite image photographed in multiple bands on an observing satellite, and verifies a multiple regression model every time a satellite image is additionally input. By updating the multiple regression model, the accuracy of the multiple regression model is gradually improved and the multiple regression model is acquired or updated after the bands are selected according to the correlation analysis and statistical analysis. Thus, the water quality monitoring ≪ / RTI >

The method of monitoring the water quality of a river system, a river system, a lake system, etc., generally adopts a method of directly sampling a sample at a specific point in the water system, but this point sampling and direct analysis The method can obtain the water quality of a certain point accurately, but can not monitor the water quality of the whole water system.

The reason why water quality should be monitored for the entire water system is because it is necessary to analyze the distribution of pollutants, the movement route, and the movement speed in various ways.

As a conventional technique for monitoring the quality of water throughout the water system, Japanese Patent Application Laid-Open No. 10-2011-0067964 proposes a method of monitoring the water quality by analyzing an image taken by a satellite, which is a satellite. In the above-described conventional art, water quality data is obtained for an area of a water body that is not measured by using a correlation formula between satellite image data and actual data in multi-bands.

However, the above-mentioned prior art uses a correlation formula that uses a specific band as an independent variable for each water quality item, and thus has a problem that the performance depends largely on the quality of the satellite image and the form of the correlation formula. That is, since the quality of the satellite image fluctuates largely due to various factors such as the shooting environment and the weather conditions, the specified correlation correlation data and data alone can not quantitatively show the water quality over the entire water system, and the correlation formula is fixed for each water quality item The error of the water quality analysis becomes very large when there is an error of the correlation formula itself.

KR 10-2011-0067964 A Jun 26, 2011.

Accordingly, an object of the present invention is to provide an accurate correlation model by correlated analysis and statistical analysis for monitoring water quality across a water system using a correlation model between a satellite image and water quality data, The present invention provides a water quality monitoring method using an observation satellite image that can accurately monitor water quality even when using a satellite image that changes.

It is another object of the present invention to provide a water quality monitoring method using an observation satellite image that minimizes an error caused by a correlation model itself by continuously updating the correlation model without specifying the correlation model.

In order to achieve the above object, the present invention provides a water quality monitoring method for receiving and analyzing satellite images of an aquatic system to provide water quality information, the method comprising the steps of: A basic data collection step (S10) of receiving the water quality data measured in accordance with the shooting date of the satellite image at different locations by different dates; After image preprocessing including Geometric Correction and Radiometric Correction is performed on the satellite image per band, the reflectance of each band is acquired and the satellite image, water quality data and reflectivity are stored in the database (S20); A band is selected by analyzing the correlation between the water quality data of the designated point and the reflectance of each band corresponding to the position of the designated spot, a multiple regression model in which the selected band reflectivity is set as an independent variable and the water quality is a dependent variable Acquiring a correlation model (S30); The satellite image of the multi-band satellite image and the water quality data of the designated spot are additionally input and subjected to image preprocessing, and the water quality data of the designated spot calculated by the multiple regression model according to the reflectance of the satellite image imposed on the additional location, If the difference is greater than a predetermined error, the process returns to the correlation model acquisition step S30 to reflect the reflectivity of the satellite image and the additional input water quality data to the previous reflectivity and water quality data to update the multiple regression model (S40); A water quality information providing step (S50) of generating a water quality map for the entire water system by substituting the reflectance of the satellite image into the multiple regression model obtained by the correlation model acquiring step (S30); And a control unit.

In the multiple regression model,

은 계수이고,

은 다중밴드 위성영상에서 얻는 밴드별 반사도이고, N은 밴드의 개수이고, 상기 지정개소에 대응되는 픽셀마다 획득한 밴드별 반사도를 독립변수로 하고, 상기 지정개소에서 측정한 수질 데이터의 수질항목을 종속변수로 한 다중회귀분석에 의해 계수를 얻어 다중회귀모델을 획득함을 특징으로 한다.Where Y is a water quality item,

Is a coefficient,

Wherein N is the number of bands, and the reflectance of each band obtained for each pixel corresponding to the designated point is used as an independent variable, and the water quality item of the water quality data measured at the designated point is expressed as And a multiple regression model is obtained by taking a coefficient by a multiple regression analysis as a dependent variable.

The correlation model acquiring step S30 may be performed before the multiple regression model is obtained by calculating a coefficient of correlation for each water quality item by the reflectivity of each band to obtain a band having a correlation coefficient greater than a preset threshold value And a coefficient of the remaining bands except for the selected band is set to '0', and the multiple regression model is obtained.

In the correlation model acquisition step S30, when the multiple regression model is updated according to the multi-band satellite image and the water quality data at the designated location, which are further input in the verification step S40, When the number of the satellite images is equal to or more than the predetermined number, the satellite image is compared with the error between the water quality value calculated by the multiple regression model and the stored water quality data, And a plurality of regression models are acquired using only one satellite image.

The water quality item includes at least one of water temperature (Temp), chlorophyll-a (Chl-a), suspended solids (SS) and total phosphorus concentration (TP).

According to the present invention configured as described above, a multiple regression model is created based on the correlation between the water quality and the reflectivity of each band, and a multiple regression model is generated for each input of the satellite image and the actual data It is possible to improve the accuracy of the multiple regression model statistically even if there is an error due to the correlation model error and the inaccuracy of the satellite image and to provide the water quality information using the multiple regression model update and the multiple regression model It is possible to construct a system that is gradually improved according to the accumulation of basic data.

1 is a block diagram of a water quality analysis apparatus for implementing a water quality monitoring method using an observation satellite image according to an embodiment of the present invention.
2 is a flowchart of a method for monitoring water quality through an observation satellite image according to an embodiment of the present invention.
FIG. 3 is an exemplary view showing a water quality distribution chart for chlorophyll-a and suspended matter provided by a method for monitoring water quality through an observation satellite image according to an embodiment of the present invention. FIG.
FIG. 4 is an exemplary view illustrating a water quality distribution map for a total phosphorus concentration and water temperature provided by a method of monitoring water quality through an observation satellite image according to an embodiment of the present invention; FIG.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings. It should be noted that, in the drawings, the same reference numerals are used to denote the same or similar components in other drawings. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS In the following description of the present invention, detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear.

FIG. 1 is a block diagram of a water quality analyzing apparatus according to an embodiment of the present invention. FIG. 2 is a block diagram of a water quality analyzing apparatus according to the present invention. Is a flowchart.

1, a water quality analyzing apparatus for implementing a water quality monitoring method using an observation satellite image according to an embodiment of the present invention includes a data collecting unit 10, an image preprocessing unit 20, (30), a water quality information providing unit (40), and a database unit (50).

The data collection unit 10 receives a satellite image obtained by photographing a water system in multiple bands from an observation satellite and water quality data measured in accordance with a photographing date of each satellite image at a plurality of designated locations in the water system In order to obtain the initial multiple regression model, satellite images and water quality data acquired for different dates are input. As described later, the multi-band satellite image of the recent date is input after the multiple regression model is initially obtained, and the received multi-band satellite image is decoded to calculate and analyze the water quality using the multiple regression model, The water quality data measured at the designated place on the date is also input, and the multiple regression model is verified as described later and updated according to the verification result.

Here, the designated location is a water quality inspection site generally performed in order to monitor the water quality of the water system, and the water quality is periodically designated by the related organization, but the water quality map for the entire water system is not shown. According to the present invention, a water distribution map for the whole water system is created by a multiple regression model as described later.

The image preprocessing unit 20 corrects a distorted image of an inputted multi-band satellite image and generally includes a geometric correction 21 and a radiation correction 22. [

The Geometric Correction 21 corrects the geometrically distorted image by the attitude of the satellite, the curvature of the earth, the direction of the satellite, the error of the observation instrument, the earth rotation, To transform the coordinates by creating a relational expression between the position in the image and the position on the actual ground coordinates. Normally, the ground reference point should be at least three or more, and the greater the number, the more accurate the correction can be.

In the embodiment of the present invention, the histogram of the surroundings is analyzed to remove errors due to atmospheric absorption and scattering, so that the minimum brightness value After determining the brightness value, a histogram djustment method was adopted to reduce the error caused by the atmosphere.

In the embodiment of the present invention, only the geometric correction 21 and the radiation correction 22 are exemplified, but it is obvious that various correction methods for correcting the distortion of the satellite image can be added.

The correlation model obtaining unit 30 is a component for obtaining a correlation model composed of a multiple regression model for estimating the water quality from the DN (Digital Number) value of each pixel of the satellite image. The correlation model obtaining unit 30 includes a DN extracting unit 31, A correlation analysis unit 33, a multiple regression model acquisition unit 34, and a satellite image sorting unit 35. [0031]

The DN extracting unit 31 extracts a DN (Digital Number) value of a pixel for each band corresponding to the position of the designated point where the water quality is measured. In order to overcome the resolution limit of the satellite image and ensure the accuracy of the DN value, it is preferable to cover the window including the peripheral pixels (for example, 3X3 pixels) of the pixel corresponding to the designated point to take an average value.

The reflectivity obtaining unit 32 is a component that converts a band-specific DN value for a designated portion into a reflectance. The reason for converting the DN value to reflectivity is that the DN image is distorted due to the atmospheric influence of the satellite image captured by the observation satellite on different dates or by using different sensors, Needs to be. In particular, since the sun altitude angle and the optical characteristics of the ground surface change depending on the shooting time of the satellite image, the solar altitude angle must be corrected to accurately ascertain the optical characteristics of the ground surface, thereby increasing the measurement reliability.

In order to convert the DN value of the satellite image into the reflectance, the DN value is firstly converted into radiance, and the reflectance can be obtained from this degree of radiance. The conversion of the DN value and the conversion of the reflectance are as follows It is made up of formulas.

here,

: 방사도

: Emissivity

: DN 값이 0일 때의 방사도

: Emissivity when the DN value is 0

: DN 값이 255일 때의 방사도

: Emissivity when the DN value is 255

: 밴드별 DN 값 중에 최고치(=255)

: Maximum value (= 255) among DN values per band

: DN 값

: DN value

.

here,

: 반사도

: Reflectivity

: 방사도

: Emissivity

d: Distance between Earth and Sun

: 외기권에서의 방사도

: Emissivity in outer space

: 태양 천정각

: Solar zenith angle

.

The reflectance corresponding to the DN value of the designated portion can be obtained for each band by using Equations (1) and (2).

The correlation between the reflectance obtained from the satellite image and the water quality data obtained from the measured data on the ground is high for the designated spot, so that water quality data can be reliably estimated from the reflectivity of the satellite image.

Accordingly, the correlation analyzing unit 33 analyzes which of the multiple bands of the satellite image has correlation with the water quality data, selects a band in which the correlation is relatively high, and sets the reflectivity of the selected band as an independent variable Is obtained by the multiple regression model acquiring unit 34 as will be described later.

In the embodiment of the present invention, to analyze the correlation between the band and the water quality data, a coefficient of correlation coefficient for each band is calculated, and a band whose correlation coefficient is larger than a preset threshold value is selected. For the bands not selected here, we obtain a multiple regression model with only the selected band as a dependent variable with a coefficient of '0' in the multiple regression model described later. On the other hand, since the water quality data measured at a plurality of designated points on the same day are used, correlation coefficients between the water quality data measured at the designated points and the reflectances corresponding to the designated points are calculated for each day, After the correlation coefficient is normalized again (or multiple times), the band is selected by comparing the standardized correlation coefficient with a preset threshold value.

In the embodiment of the present invention, water quality data including water temperature (Temp), chlorophyll-a (Chl-a), suspended solids (SS) and total phosphorus concentration (TP) And therefore, the correlation coefficient is calculated for each of the water quality items, and bands are selected for each water quality item.

As described above, the multiple regression model acquiring unit 34 derives a multiple regression model in which the degree of reflectivity of a band in which correlation is relatively high is an independent variable and the water quality item is a dependent variable, and the basic model of a multiple regression model (3). &Quot; (3) "

은 계수이고,

은 다중밴드 위성영상에서 얻는 밴드별 반사도이고, N은 밴드의 갯수이다.Here, Y is a water quality item,

Is a coefficient,

Is the reflectivity of each band obtained from a multi-band satellite image, and N is the number of bands.

In addition, the multiple regression model is a multiple regression analysis in which only bands selected with the coefficients of the remaining bands as '0' are independent variables and water quality items are dependent variables except the bands selected by the water quality item as described above , Multiple return analysis).

Exceptionally, since the water temperature has a high correlation with the band of the infrared infrared wavelength, it is preferable to obtain the multiple regression model with the band selected in advance for the water temperature. .

On the other hand, in order to improve the accuracy of the multiple regression model, it is desirable to select a suitable satellite image and acquire a multiple regression model using only the selected satellite image. Since the DN value of the satellite image is corrected by the image ionizer 20 and converted into the reflectance, errors due to various environmental factors can not be completely eliminated. Therefore, when a multiple regression model is obtained from a plurality of satellite images, . For example, if the weather condition is not very good and the error is large, it is better to obtain a multiple regression model with only satellite images that exclude such satellite images. Therefore, as shown in FIG. 1, the correlation model obtaining unit 30 further includes a satellite image sorting unit 35. FIG.

If the number of satellite images to be used in updating the multiple regression model is more than a predetermined number and the satellite image and water quality data are additionally input as described later, The water quality is calculated for each satellite image by using the multiple regression model obtained by the regression model acquiring unit 34 (that is, the water quality is calculated by substituting the reflectance of the pixel corresponding to the designated point) And selects a satellite image having a preset error or less. Then, the multiple regression model acquiring unit 34 modifies the multiple regression model with only the selected satellite image. Of course, since the satellite image, the water quality data, and the reflectivity for the designated location are stored and managed in the database unit 50, the water quality data can be calculated using the current multiple regression model based on the past satellite images.

By selecting the satellite images in this way, it is possible to improve the accuracy of the multiple regression model using the selected satellite images rather than the multiple regression models before selecting the satellite images.

The water quality information providing unit 40 creates a water quality map by applying a multiple regression model for each water quality item obtained from the correlation model obtaining unit 30 to a specific satellite image, To consumers. Here, the specific satellite image refers to a satellite image received every time when the satellite image is received continuously, as can be seen from the flowchart of FIG. It is also apparent that the water quality map for the specific satellite image can be stored in the database unit 50 to provide a water quality map for the satellite image of the requested date by the related authority.

The water quality map is composed of a reflectivity obtaining unit 41 for obtaining the reflectance for each band in the entire satellite image, a masking file creating unit 42 for masking only the water portion in the specific satellite image to obtain the water image, And the water quality distribution map for each water quality item, which is displayed in the water-based image obtained by calculating the calculated water quality in pixel units, by assigning the band-by-band reflectance to the multiple regression model for each water quality item 43).

Here, it is preferable that the mysacking file creating unit 42 sets in advance a satellite image of a band that is easy to obtain a water image from the satellite image of each band. And, it is good to display the water quality by dividing the water quality chart by color.

On the other hand, the water quality information providing unit 40 further includes a verification unit 44.

The verification unit 44 further receives a multi-band satellite image for creating a water quality map and further receives the water quality data of the designated location measured at the photographing date of the multi-band satellite image, and the reflectivity acquiring unit 41, The masking file creation unit 42 and the water quality map generation unit 43 are operated to obtain the water quality data in the process of providing the water quality map corresponding to the satellite image additionally inputted. Compare with the data. When the error is larger than a preset error, the correlation model obtaining unit 30 is operated to update the multiple regression model. In this case, the multiple regression model update is applied by applying the multi-band satellite image and the water quality data together with the previous multi-band satellite image and the water quality data to obtain a multiple regression model reflecting the additional input data.

The database unit 40 stores the satellite image and the water quality data collected by the data collection unit 10, the reflectivity and the multiple regression model generated by the correlation model obtaining unit 30, And the obtained water quality distribution map. The water quality data and reflectivity thus stored and reflected are reflected together with the reflectivity of the water quality data to be additionally input and the satellite image to be additionally received when the multiple regression model is updated. And, the water quality distribution map to be stored and managed is transmitted through the communication network to be utilized by the related organization for the water quality map of the requested date. Of course, the water quality map can also be provided in the form of a bulletin board through the Internet.

Hereinafter, a method for monitoring water quality through an observation satellite image according to an embodiment of the present invention will be described with reference to FIG. 2, and redundant description will be omitted with reference to the block diagram of FIG.

According to an embodiment of the present invention, in order to obtain an initial multiple regression model, a basic data collection step (S10) for collecting basic data including multi-band satellite images and water quality data, a multi- (S30) for acquiring a multiple regression model based on reflectivity and water quality data (S30), a verification step (S40) for verifying multiple regression models, and a water quality using a multiple regression model And a water quality information providing step (S50) for providing a distribution chart.

The basic data collection step (S10) is a basic data for obtaining an initial multiple regression model for each water quality item. In the multi-band satellite image taken by the observation satellite and a plurality of designation points of the water system, And receiving the measured water quality data through the data collection unit 10 for different dates.

According to the present invention, an initial multiple regression model is acquired, and water quality information based on a satellite image is provided, and a new multiple satellite image is updated every time a new satellite image is received to update a multiple regression model. In order to obtain the initial multiple regression model, it is necessary to obtain the data acquired from a plurality of dates, since a plurality of satellite images photographed on different dates are required and water quality data measured from the ground is required for each date.

The water quality data of the designated location collected in the basic data collection step S10 includes at least one of water temperature Temp, chlorophyll-a (Chl-a), suspended solids SS and total phosphorus concentration TP It is the data as the water quality item.

In the image processing step S20, the multi-band satellite images collected as basic data are subjected to image preprocessing (including geometric correction and radiation correction) in the video pre-processing unit 20 to correct distortions (S21) The DN extraction unit 31 and the reflectivity acquisition unit 32 of the unit 30 acquire the reflectance of each date and band by the designated portion (S22). At this time, the reflectance by date and by band is the reflectance of the pixel corresponding to the acquisition point of the water quality data inputted as the basic data, that is, the position of the plurality of designated points.

The reflectance of each point, date, and band obtained in the image processing step S20 is stored in the database unit 50 together with the water quality data and the satellite image input in the basic data collection step S10 . At this time, the reflectivity and the water quality data to be stored are used again when the multiple regression model is updated by the following verification step (S40).

The correlation model obtaining step S30 is a step of obtaining a multiple regression model by the correlation analyzing unit 33 and the multiple regression model obtaining unit 34 of the correlation model obtaining unit 30, A correlation coefficient between the reflectivity of each of the dates and bands corresponding to the designated spot obtained in step S20 and the water quality data received in the basic data collection step S10 is calculated to select a band in which the correlation coefficient is higher than a preset threshold value (S31), multiple regression analysis is performed on the selected band reflectivity and input water quality data, and a multiple regression model is obtained in which a band having a high correlation is an independent variable and a water quality item is a dependent variable (S32). Here, the selection of the high correlation band and the acquisition of the multiple regression model are performed separately for each water quality item.

The following table is a multiple regression model obtained by applying to a lake formed in any dam in Korea as a concrete example. The multi-band satellite images are seven band images taken by a TM sensor mounted on LANDSAT, and three identical images taken on different dates were adopted. The water quality data are chlorophyll-a (Chl-a), floating The measured data at six different locations are applied to the water quality items (SS), total phosphorus concentration (TP) and water temperature (Temp).

Water quality items Selected band Multiple regression model Determination coefficient
(R ² ) Chl-a B ₂ , B ₃ , B ₄ Chl-a = -44 + 2929.6B ₂ + 1062.7B ₃ + 626B ₄ 0.79 SS B ₂ , B ₃ SS = -10.5 + 446.4B ₂ + 685.1B ₃ 0.67 TP B ₂ , B ₃ TP = -0.03 + 1.92B ₂ + 2.18B ₃ 0.71 Temp B ₆ Temp = 92.39 B ₆ -51.49 0.89

Referring to Table 1, the selected bands calculated by the correlation coefficient are different for each water quality item. The multiple regression model obtained for each water quality item has the selected band as the dependent variable according to the correlation coefficient. That is, by setting the coefficient of the band that is not selected because the correlation is low or not selected to be "0" in comparison with the basic model of Equation (3), a multiple regression model in which only the coefficient of the selected band is a dependent variable is completed will be.

The decision coefficients were calculated separately for the accuracy of the multiple regression model. As shown in Table 1, the multiple regression model in which the number of satellite images and water quality data applied for the result of Table 1 are not so small and the coefficient of determination is low can be seen. However, if the multiple regression model is continuously updated based on the new satellite image and water quality data according to the following verification step (S40), a more accurate model can be obtained.

In the verification step S40, while the water quality information providing unit 40 provides the water quality information including the water quality map, when the new satellite image and the water quality data are additionally received by the data collecting unit 10, ), And then verifying.

That is, in the verification step S40, the multi-band satellite image and the water quality data of the designated point are additionally input (S41), preprocesses the image (S42), calculates the reflectivity (S43) The reflectivity is extracted (S44) and the water quality data for each water quality item at the designated location is calculated by a multiple regression model. If the calculated water quality data is compared with the input water quality data and the error is greater than a preset error, the flow returns to the correlation model acquisition step (S30) to update the multiple regression model. At this time, the update of the multiple regression model reflects both the reflectivity and the water quality data previously stored in the database unit 50 and the reflectivity and the water quality data according to the additional input, and the derivation of the correlation (i.e., correlation coefficient) The selection of bands and the acquisition of multiple regression models are the same, except that the reflectivity and water quality data are increased from the previous one.

When the multiple regression model is updated by the verification step (S40), the correlation coefficient is again calculated to select a band. In other words, the previously selected bands may have been erroneously selected due to lack of data or inaccuracies in the data (for example, errors in satellite images or correction errors) Regression model.

Therefore, according to the present invention, when the additional data is input, the previously obtained multiple regression model is verified. When the additional data is different from the previous data, the multiple regression model is renewed, Can be improved.

In order to further improve the accuracy of the multiple regression model, the satellite image selector 35 is used. That is, when the additional data is input and received together with the water quality data and the number of the satellite images to be stored and managed exceeds a predetermined number, the error between the water quality value calculated by the multiple regression model and the water quality data Compare them. Then, the satellite images whose errors are less than a preset error are selected, and a multiple regression model is again obtained using only the selected satellite images.

On the other hand, the reflectivity calculated in the verification step (S40) is described for the entire water system. This is for obtaining the water quality distribution map, which will be described later, by calculating the reflectivity with respect to the entire water system. If only the reflectance of the designated site where the water quality is measured is calculated, the reflectivity of the entire water system should be calculated in the water quality information providing step (S50), which will be described later.

The water quality information providing step (S50) is a step of generating and providing a water quality map for the entire water system by substituting the reflectivity obtained from the satellite image into the multiple regression model obtained in the correlation model acquiring step (S30) Similarly, a masking file is created to obtain an image of the entire water system (S51), and the water quality data is displayed in accordance with the pixels to create a water quality map (S52).

Since the multiple regression model at this time is preferable to adopt the updated model if it is updated and it is necessary to reflect the whole area of the water system, as described above, in the verification step (S40), the satellite image, Reflectivity is used to calculate the reflectivity for the whole water system separately for the initial data that does not calculate the reflectivity for the entire water system.

Of course, it is also good to be able to manage the reflectivity of the entire water system in the database, or to store the water quality map created for each satellite image in the database, and to provide the water quality map of the date requested by the relevant organization. It is also good to modify the water quality map using the newly updated multiple regression model for the previous satellite images.

3 and 4 show the water quality map for the water system showing the results of Table 1 above.

FIGS. 3 and 4 show the distribution of water quality by date with respect to water temperature (Temp), chlorophyll-a (Chl-a), suspended solids (SS) and total phosphorus concentration (TP) So that the water quality distribution can be recognized at a glance.

In describing the embodiment of the present invention, the multi-band satellite image is used to acquire the multiple regression model and provide the water quality information. However, instead of the satellite image, the aerial image obtained by photographing the multi- Can be used.

Since the aerial image is distorted by the shooting altitude, the shooting angle, the moving speed, the atmospheric condition, etc., the image preprocessing is performed as described above and the reflectivity is calculated. It is obvious in the technical field that the image preprocessing at this time can be performed in a manner suitable for aerial image.

Accordingly, the present invention should be considered as falling within the scope of the present invention even if an aerial image is used instead of a satellite image.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, . ≪ / RTI > Accordingly, such modifications are deemed to be within the scope of the present invention, and the scope of the present invention should be determined by the following claims.

10: data collecting unit 20: image preprocessing unit
30: correlation model acquisition unit 40: water quality information providing unit
50:

Claims (5)

A water quality monitoring method for receiving and analyzing satellite images of a water system to provide water quality information,
(S10) a basic data collection step (S10) of receiving, by different dates, water quality data measured in accordance with a shooting date of a satellite image at a plurality of designated points of a multi-band satellite image and a water system in which a water system is photographed by an observation satellite;
After image preprocessing including Geometric Correction and Radiometric Correction is performed on the satellite image per band, the reflectance of each band is acquired and the satellite image, water quality data and reflectivity are stored in the database (S20);
A band is selected by analyzing the correlation between the water quality data of the designated point and the reflectance of each band corresponding to the position of the designated spot, a multiple regression model in which the selected band reflectivity is set as an independent variable and the water quality is a dependent variable Acquiring a correlation model (S30);
The satellite image of the multi-band satellite image and the water quality data of the designated spot are additionally input and subjected to image preprocessing, and the water quality data of the designated spot calculated by the multiple regression model according to the reflectance of the satellite image imposed on the additional location, If the difference is greater than a predetermined error, the process returns to the correlation model acquisition step S30 to reflect the reflectivity of the satellite image and the additional input water quality data to the previous reflectivity and water quality data to update the multiple regression model (S40);
A water quality information providing step (S50) of generating a water quality map for the entire water system by substituting the reflectance of the satellite image into the multiple regression model obtained by the correlation model acquiring step (S30);
And monitoring the quality of the water through the observation satellite image. The method according to claim 1,
In the multiple regression model,

Where Y is a water quality item,

Is a coefficient,

Is the reflectivity of each band obtained from the multi-band satellite image, N is the number of bands,
A multiple regression model is obtained by taking a coefficient by a multiple regression analysis in which the water level data of the water quality data measured at the designated point is used as a dependent variable and the reflectivity of each band corresponding to the designated point is taken as an independent variable A method of monitoring water quality through observation satellite images. 3. The method of claim 2,
The correlation model acquisition step (S30)
Prior to obtaining the multiple regression model, a band having a correlation coefficient greater than a predetermined threshold value is calculated by calculating a coefficient of correlation (coefficient of correlation) between the band-specific reflectance and the water quality item, and the remaining band excluding the selected band And the multiple regression model is obtained after setting the coefficient of the water quality monitoring unit to '0'. The method of claim 3,
The correlation model acquisition step (S30)
When the multiple regression model is updated according to the multi-band satellite image and the water quality data of the designated point, which are further inputted in the verification step (S40), the correlation coefficient is again calculated to select a band,
When the number of the satellite images is equal to or more than a predetermined number, the satellite image is compared with the error between the water quality value calculated by the multiple regression model and the stored water quality data, The method of monitoring water quality through observation satellite images is characterized in that a plurality of regression models are acquired and used. 5. The method of claim 4,
The water quality item includes:
Wherein at least one of water temperature (Temp), chlorophyll-a (Chl-a), suspended solids (SS) and total phosphorus concentration (TP) is included.

Images