KR20110094902A - Normalized difference vegetation index correction method using spatio-temporal continuity - Google Patents

Abstract

PURPOSE: A vegetation index correction method using visual continuous is provided to supply a good quality of vegetation index. CONSTITUTION: An MODIS(Moderate Resolution Imaging Spectroradiometer) vegetation index information and quality information are obtained from a vegetation index generation center(S300). A correction area is determined among the MODIS vegetation index information(S310). The MODIS vegetation index information project the correction area through an MRT(MODIS Reprojection Tool) and is attached to the correction area as a material.

Description

Normalized difference vegetation index correction method using spatio-temporal continuity}

The present invention relates to a method for calibrating vegetation index (or simply referred to as vegetation number) using space-time continuity, which is caused by cloud or snow pollution and missing, aerosol, etc. in continuous vegetation index information generated from remote sensing using satellite. The present invention relates to a vegetation index correction method that eliminates problems such as atmospheric effects and exploration noise.

The vegetation index is used as representative information indicating the vegetation state of the ground. Despite the importance of this vegetation index, it is practically impossible to construct regional information through field observations. As an alternative, a wide range of vegetation indices are used to calculate the wide vegetation index.

However, vegetation index information generated from remote sensing information generated by such satellites not only has distorted information due to cloud pollution or missing or influence of the atmosphere, but also has a problem in that it is not continuous in time and space.

The present invention has been proposed to solve the problems according to the prior art, and an object thereof is to provide a correction method for removing distorted information from vegetation index information generated by a satellite.

Another object of the present invention is to provide a correction method for correcting vegetation index information generated by a satellite with space-time continuity.

The present invention provides a vegetation index correction method using space-time continuity, in order to achieve the problem posed above. The vegetation index correction method includes: an acquisition step of obtaining MODIS (Moderate Resolution Imaging Spectroradiometer) vegetation index information and quality information of raw data from a vegetation index generation center; Reprojection step of attaching the MODIS vegetation index information as one data to the correction region by determining a correction region of the MODIS vegetation index information and reprojecting using an MDIS (MODIS Reprojection Tool); A synthesis step of synthesizing monthly vegetation index information for at least one year from the MODIS vegetation index information based on the quality information, based on only the upper quality pixels; A waterside masking step of masking watersides distinguishing land and watersides from the monthly vegetation index information of the at least one year; Determining pixel defects by checking temporal discontinuity of masked monthly vegetation index information by at least three months to determine whether at least one monthly pixel is defective, and correcting the monthly vegetation index information; An available pixel checking step of checking an available pixel within a radius R of a specific area of the correction area; A usable pixel determination step of determining whether there are five or more usable pixels and correcting the monthly vegetation index information; A remaining pixel defect determining step of inspecting all the pixels in the correction region to determine whether pixel defects remain; And an iteration step of repeating the waterside masking step or the remaining pixel defect determination step by increasing the radius R. FIG.

In this case, the pixel failure determination step, if the at least one month-by-month pixel is bad, the arithmetic mean of the previous month vegetation index value and the second month vegetation index value of the at least three months to the middle month vegetation midway between the previous month and the month It may include the step of correcting to an index value.

At this time, the water masking step, L / W (Land / Water) by using the master leaf area index (LAI) & Land Cover (LC) information included in the MODIS vegetation index information of the raw data (Land / Water) May include masking the land / waterfront).

In this case, the available pixel determination step, if there are five or more available pixels, the following equation

, w는 가중 평균이고, d는 현재 화소를 중심으로 각 화소까지의 거리를 나타냄)을 이용하여 상기 월별 식생 지수를 보정하는 단계를 포함할 수 있다. (here,

, w is a weighted average, and d represents a distance from each pixel to the center of the current pixel).

In this case, the determining of the residual pixel defect may include increasing a radius R of the specific region.

을 이용하여 상기 수변중 해안선의 손실을 보상하는 단계; 및

을 이용하여 상기 육지중 도시 및 사막 등의 지역의 손실을 보상하는 단계를 더 포함할 수 있다. At this time, the repeating step,

Compensating for the loss of the shoreline under the water; And

Compensating for the loss of the land, such as cities and deserts on land may be further included.

In this case, the MDIS (MODIS Reprojection Tool) may use a Plate Carree map projection method.

According to the present invention, the quality of the vegetation index information and ground cover classification information is corrected using the vegetation index information to remove the pollution and missing due to cloud or snow, the effect of the atmosphere by aerosol, etc. to improve the high quality vegetation index can do.

Another effect of the present invention is that the vegetation index having a continuity in time can be calculated by calculating the arithmetic mean or the weighted average from the vegetation index information of consecutive periods.

1 is a conceptual diagram illustrating a network implementing vegetation index correction showing a process of obtaining vegetation index information according to an embodiment of the present invention.
2 is a circuit block diagram of the computer system of FIG.
3 is a flowchart illustrating a process of correcting obtained vegetation index information according to an embodiment of the present invention.
4 is a screen example for designating a correction area according to an embodiment of the present invention.
5 is a diagram illustrating an example of quality information provided with MODIS vegetation information according to an embodiment of the present invention.
6 is a graph showing the percentage of defective defective pixels of the vegetation index after each correction step according to an embodiment of the present invention.
7 is a screen example before and after the correction of the vegetation index according to an embodiment of the present invention.
8 is a graph showing a vegetation index of the monthly average according to an embodiment of the present invention.

Hereinafter, a method of correcting a vegetation index obtained by a satellite according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

1 is a conceptual diagram illustrating a network implementing vegetation index correction showing a process of obtaining vegetation index information according to an embodiment of the present invention. This network is based on satellites 110 that search the Earth's surface, and information derived from the satellites 110, such as the Master Leaf Area Index, Land Cover, and Vegetation Index (NDVI). A vegetation index generation center 120 for generating Vegatation Index) information, and a computer system 140 connected to the vegetation index generation center 120 through a communication network 130 to correct the vegetation index. These network components are described as follows.

The satellite 110 may be, for example, a meteorological satellite equipped with a MODIS (Moderate Resolution Imaging Spectroradiometer). The satellite 110 traverses the earth 100 in a low orbit and searches for the surface of the earth.

The vegetation index generation center 120 generates data such as MODIS vegetation index, master leaf area index (LAI), ground cover (LC), etc. using information about the earth 100 searched by the satellite 110, It is responsible for distributing these data by wire or wirelessly.

Of course, there is a vegetation index such as GIMMS (Global Inventory Modeling and Mapping Studies), SPOT Vegetation, etc., but in one embodiment of the present invention will be described as MODIS vegetation index for convenience of understanding.

Here, the ground cover refers to information classified and classified according to each geographical position of the ground surface composition of the earth 100. The type depends on which criteria are classified by the International Geosphere Biosphere Program (IGBP), the University of Mayland (UMd), the US Geological Survey (MOGS), or the MODIS, or by what purpose.

In an embodiment of the present invention, although USGS, UMd, or IGBP is taken as an example, there may be numerous ground cover classification information. The types of classification of this information are cities, farmland, evergreen coniferous trees, evergreen deciduous forests, deciduous deciduous forests, deciduous coniferous forests, savannas and grasslands.

In other words, information that separates and bundles pixels having the same or similar ground characteristics becomes the surface covering classification information.

The master leaf area index and ground cover classification information uses the information contained in the background information of the Meteorological Agency's next-generation numerical forecast model called UM (Unified Model).

The computer system 140 is connected to the vegetation index generation center 120 through the communication network 130, and is a device capable of receiving and correcting the vegetation index generated by the vegetation index generation center 120. Therefore, the computer system 140 requires a configuration for correcting the vegetation index according to an embodiment of the present invention. A diagram showing this is shown in FIG. 2. 2 will be described later.

Here, the communication network 130 is a wired Internet network using a wired communication technology such as HTTP, or WiBro (Wireless Broadband), WiPi (Wireless Internet Platform for Interoperability), Bluetooth, Code Division Multiple Access (CDMA), IrDA. It may be a wireless network using a wireless communication technology such as (Infrared Data Association).

The computer system 140 may be configured with a database 141, which stores master LAI & LC information, MODIS vegetation index information, and the like. Of course, such a database 141 may be implemented on the computer system 140, but a standalone system is possible only for the database. For example, it could be a database server.

2 is a circuit block diagram of the computer system of FIG. Referring to FIG. 2, a circuit block diagram of a computer system includes a comparison unit 220 comparing a key input unit 210, master LAI & LC information, MODIS vegetation index information, and the like to find a difference, a correction unit 230 correcting according to the difference, A storage unit 240 for storing an algorithm for comparing the MODIS vegetation index information and the like to find and correct the difference, a display unit 250 for displaying information, and a communication unit 250 for communicating with the outside are configured. These components are described as follows.

The controller 200 controls the computer system 140 as a CPU (Central Processing Unit), and exchanges signals and data with other components.

The key input unit 210 serves to transmit a user's command to the computer system 140 as an input means such as a keyboard or a mouse.

The comparator 220 compares the quality information of the MODIS vegetation index information and the master LAI & LC information of the specific region transmitted through the communication unit 260 in time and space to find a difference.

The corrector 230 corrects the difference found by the comparator 220.

The storage unit 240 is composed of flash memory, electrically erasable programmable read only memory (EEPROM), magnetoresistive random access memory (MRAM), ferroelectric RAM (FeRAM), phase change RAM (PRAM), SRAM, DRAM, and hard disk drives. The controller 200 may include data, algorithms, and the like for the controller 200 to operate.

Of course, this algorithm is implemented by program software, and compares the quality information of MODIS vegetation index information and master LAI & LC information of a specific area with time and space to find the difference, and corrects this difference to provide high quality vegetation with time and space continuity. Function to generate index information.

The display unit 250 serves to display an input screen, a result screen, and the like, and provide the same to a user. To this end, the display unit 250 may be configured of a liquid crystal display (LCD), an organic light emitting display (OLED), or the like, and an independent display such as a monitor may be used. Of course, in this case, a display adapter or the like for connecting the monitor and the computer system 140 is used.

The communication unit 260 functions to connect with the communication network 130 of FIG. 1, thus transmitting and receiving data wirelessly or by wire. Of course, in the case of wired, a modem (not shown) is provided, and in the case of wireless, a wireless communication circuit is configured for communication.

Next, a process of correcting the obtained vegetation index information will be described with reference to FIGS. 1 and 2. An example illustrating this is illustrated in FIG. 3, and FIG. 3 is a flowchart illustrating a process of correcting acquired vegetation index information according to an embodiment of the present invention.

Referring to FIG. 3, first, the computer system 140 obtains vegetation index information from the vegetation index generation center 120 through the communication network 130 (step S300). In detail, a shell script for the ftp (file transfer protocol) protocol may be configured to selectively download MODIS vegetation index information corresponding to a predetermined research area through the communication network 130, for example, National Aeronautics and NASA. Space Administration) can be downloaded from the ftp server (120 of FIG. 1).

Of course, MODIS vegetation index information is provided by dividing data by a certain area of granule unit.

When the MODIS vegetation index information is obtained, an area to be corrected is determined from the MODIS vegetation index information (step S310). In other words, this step is a process of determining an area for performing the correction, and FIG. 5 shows an example of the screen for easy understanding. That is, FIG. 5 is a screen example in which the minimum correction area 400 requiring correction is expressed in a fan shape together with latitude and longitude.

This correction area 400 includes a minimum range and can be used in future corrections and specified from 70 degrees to 180 degrees east and 90 degrees north to 60 degrees south to minimize geographic discontinuity in each correction. Select the area.

4 shows an example of a fan-shaped correction region 400 in the clockwise direction of 76.102E 53.061N, 176.114E 52.984N, 154.049E 10.36N, 98.098E 10.415N.

Of course, this is for the convenience of understanding the present invention, and the correction area 400 does not have to be fan-shaped, and may be rectangular or circular, and a plurality of correction areas may be designated.

If the correction area is determined, it is reprojected using an MDIS (MODIS Reprojection Tool) (step S320). In other words, MRT is a USDIS-provided MODIS data reprojection tool, which reprojects MODIS vegetation index information selected as a correction region from sinusoidal projection to user desired map projection (eg plate carree map projection). .

That is, in order to make the handling of MODIS vegetation index information convenient, in one embodiment of the present invention, the MODIS vegetation index information divided into granules by using MRT is pasted as one data for the previously set area, and at the same time, the desired area by map projection method. Reproject. Of course, sine curve projection as well as plate carree map projection can be reprojected in several ways.

An example showing the quality information of the MODIS vegetation index information generated by this reprojection process is shown in FIG. 5. That is, FIG. 5 is a diagram illustrating an example of quality information provided with MODIS vegetation information according to an embodiment of the present invention. An example 500 of dual quality information is as follows.

[0-1]: MODLAND_QA [2 bit range]

 00: VI produced, good quality

 01: VI produced, but check other QA

 10: Pixel produced, but most probably cloudy

11: Pixel not produced due to other reasons than clouds

In detail, FIG. 5 illustrates an example for describing the quality of the MODIS vegetation index in bits.

When the reprojection process is completed, synthesis is performed on a monthly basis using the higher quality pixel average (step S330). In other words, there is quality information generated by the reprojection process provided with MODIS vegetation index information. In an embodiment of the present invention, the vegetation index information is collected for three years (2006-2008), and based on the quality information of each of the collected MODIS vegetation indices, the average of only higher quality information is 1 year MODIS per cycle. Calculate the vegetation index data.

In other words, each monthly (period) vegetation index information is prepared with quality information for each year (of course, created by using MRT), and from this, the average MODIS vegetation index of three years is calculated as one data. Calculate, but use only high quality data according to quality information.

Of course, the three-year time period described above is exemplary, and it is also possible to collect MODIS vegetation index information by setting a period longer than three years.

When the synthesis is made, waterside masking is performed (step S330). In other words, the Master Leaf Area Index & Land Cover (LAI & LC) data is the basic data included in the numerical model called Unified Modle (UM), and the L / W (Land / Water) is used to mask the water and land. (S342 and S344). Here, LAI is the leaf area index, LC is the ground cover, and the ground cover is the data classified and classified according to the geographic location of the surface composition.

Of course, it is possible to distinguish the waterfront only with MODIS vegetation index information, but the waterside masking step (S330) is a process of setting the waterfront to a specific value. Can be.

In addition, the shoreline region may be distinguished from the land and the waterfront (sea) classification data (step S346). In other words, because the nature of the vegetation index for this part requires a series of different correction process, it is displayed separately.

When the waterside masking step S340 is completed, the temporal discontinuity is checked (step S350). In other words, this is a process of confirming whether high quality information is composed of continuous data in time. For better understanding, for example, the March and April data are examined together for the March data, and the February and April data, except March, are checked for higher quality.

In other words, it is determined whether only one pixel is bad by comparing only the data of the current period (that is, the data in March in the example) among the above three data (that is, the data of March in the example) (step S360).

As a result of the determination, if only one pixel (i.e., data in March) is defective, an arithmetic average is made (step S361). In other words, for March vegetation index correction, February and April are cases of higher quality data. In this case, the arithmetic mean of February and April is corrected to the vegetation index value of the month, that is, March.

In contrast, since only one pixel cannot be corrected unless it is defective, available pixels in the region R (where R means the radius of the correction region 400 shown in FIG. 4) (ie, the same surface coating type and higher quality). In pixel) (step S370). In this case, R is initially given as 5, which examines 5 * 5 pixels, that is, 25 pixels in the x and y directions in space with respect to the current pixel (bad). In other words, this is a process of checking the number of pixels having the same ground coating type and higher quality as the current pixel among the 25 pixels.

As a result of the pixel check, it is determined whether there are five or more available high quality pixels (step S380). In other words, as a result of the inspection of step S370, it is a process of checking whether at least five pixels of the same surface coating type and higher quality are at least five of the 25 pixels.

As a result of determination, when there are five or more available high quality pixels, the current pixel is corrected by using the weighted average of the distances to the respective pixels as shown in the following equation (step S381).

, w는 가중 평균이고, d는 현재 화소를 중심으로 각 화소까지의 거리를 나타낸다.here,

, w is a weighted average, and d represents the distance to each pixel with respect to the current pixel.

In step S380, if there are not five or more available high quality pixels, the condition is not satisfied and the correction is not performed, and all pixels are inspected to determine whether there are still bad pixels that need correction (step S390).

As a result of the bad pixel checking, if the bad pixel remains, steps S350 to S390 which are correction processes are repeatedly performed. Here, there is a process of checking whether the correction process has been repeated three or more times (step S391). This is because, since the correction of most of the land pixels needs to be completed before the exception correction process S331 is performed, the exception correction process is performed together when the correction process is repeated at least three times.

Even if the correction process is repeated three or more times, the R value is not increased to infinity (step S391). In other words, the R value is increased from " 5 " to " 7 " and steps S350 to S390 are repeatedly performed, but the R value is not increased above " 9 " even when the correction process is repeated three or more times. This is because even if the pixels are no longer distant, the same type as the current pixel and the higher quality may be different depending on the geographical position.

On the other hand, in another embodiment of the present invention is an exception correction process (S331). The exception correction process includes a coastal loss correction step S333 and a loss correction step S335 such as a city.

In other words, the coastal loss correction step (S333) is calculated in the same manner as calculated in step S380, but only 50% of the value calculated in consideration of the geographical characteristics is applied. This is expressed as the following equation.

When coastal loss correction is made, a loss correction step S335 of a city or the like is performed. In other words, the loss correction step (S335) of the city, etc. is calculated in the same manner as calculated in step S380, but only 50% of the value calculated in consideration of the geographical characteristics is applied. This is expressed as the following equation.

The exception correction is applied after at least three corrections are repeated since the number of available pixels is only half that of the inland pixels when the correction of other land pixels is insufficient.

Next, the correction process according to FIG. 3 will be described using a screen example. 6 is a graph showing the percentage of defective defective pixels of the vegetation index after each correction step according to an embodiment of the present invention.

Referring to FIG. 6, a raw curve 600 showing the ratio of residual defective quality pixels in the uncorrected state, and a primary correction curve showing the ratio of residual defective quality pixels in the primary correction state 610, a secondary correction curve 620 showing a ratio of residual defective quality pixels in the secondary correction state, and a third correction curve 630 showing a ratio of residual defective quality pixels in the third correction state; ), A fourth-order correction curve 640 showing a ratio of residual defective quality pixels in a fourth order correction state, a fifth-order correction curve 650 showing a ratio of residual poor quality pixels in a fifth order correction state, A final correction curve 660 is shown showing the percentage of residual poor quality pixels in the final correction state.

As shown in FIG. 6, as the correction proceeds, the ratio of the residual defective quality pixels decreases.

7 is a screen example before and after the correction of the vegetation index according to an embodiment of the present invention. In other words, the screen example 700 before the correction of the vegetation index is shown on the left side, and the screen example 710 after the correction of the vegetation index is shown on the right side. In particular, Figure 7 shows the case of July, expressed in white to black according to the value of the vegetation index.

The portion indicated in green in the example screen 700 before correction means missing (pixels in which only one normal pixel does not exist). In the screen example 710 after correction, the missing green color can be seen to be completely removed.

8 is a graph showing a vegetation index of the monthly average according to an embodiment of the present invention. That is, the seasonal variation of the vegetation index for each latitude can be analyzed through the correction process, which is shown in FIG. 8. More specifically, 10N-10S (800), 30-10N (810), KOPE (Korean Peninsula) (820), 10-60S (830), Total (840), 50-30N (850), 90N-50N (860) Seasonal fluctuations of vegetation indices are continuously shown by the latitudes of, and characteristics of vegetation indices by latitude are well represented. The high vegetation index is characterized by low annual vegetation index and large annual fluctuations. Vegetation fluctuations in the mid-latitude regions, including the Korean peninsula, show relatively high annual fluctuations and the annual vegetation index is maintained at about 0.3 to 0.5.

As described above, the present invention has been described by way of limited embodiments and drawings, but the present invention is not limited to the above-described embodiments, which can be variously modified and modified by those skilled in the art to which the present invention pertains. Modifications are possible. Accordingly, the spirit of the present invention should be understood only by the claims set forth below, and all equivalent or equivalent modifications thereof will belong to the scope of the present invention.

100: Earth 110: satellite
120: Vegetation Index Generation Center
130: communication network 140: computer system
141: Master LAI & LC information storage database
200: control unit 210: key input unit
220: comparison unit 230: correction unit
240: storage unit 250: display unit
260: communication unit

Claims (7)

An acquisition step of obtaining MODIS (Moderate Resolution Imaging Spectroradiometer) vegetation index information and quality information of raw data from a vegetation index generation center;
Reprojection step of attaching the MODIS vegetation index information as one data to the correction region by determining a correction region of the MODIS vegetation index information and reprojecting using an MDIS (MODIS Reprojection Tool);
A synthesis step of synthesizing monthly vegetation index information for at least one year from the MODIS vegetation index information based on the quality information, based on only the upper quality pixels;
A waterside masking step of masking watersides distinguishing land and watersides from the monthly vegetation index information of the at least one year;
Determining pixel defects by checking temporal discontinuity of masked monthly vegetation index information by at least three months to determine whether at least one monthly pixel is defective, and correcting the monthly vegetation index information;
An available pixel checking step of checking an available pixel within a radius R of a specific area of the correction area;
A usable pixel determination step of determining whether there are five or more usable pixels and correcting the monthly vegetation index information;
A remaining pixel defect determining step of inspecting all the pixels in the correction region to determine whether pixel defects remain; And
An iterative step of repeating the waterside masking step or remaining pixel defect determination step by increasing the radius R
Vegetation index correction method using space-time continuity comprising a. The method of claim 1,
The pixel failure determination step,
If the at least one monthly pixel is defective, correcting the arithmetic mean of the previous month's vegetation index value and the last month's vegetation index value to the vegetation index value of the middle month in the middle of the previous month and the second month. Vegetation index correction method using space-time continuity. The method of claim 2,
The waterside masking step,
L / W (Land / Water: Land / Waterside) is classified by using Master Leaf Area Index (LAI) & Land Cover (LC) information included in MODIS vegetation index information of the raw data. Vegetation index correction method using space-time continuity comprising the step of masking. The method of claim 3, wherein
The usable pixel determination step,
If there are five or more of the available pixels, the following equation

(here,

and w is a weighted average, and d represents a distance to each pixel from the current pixel. 2. The method of claim 1, wherein the monthly vegetation index is corrected. The method of claim 4, wherein
The remaining pixel failure determination step,
Vegetation index correction method using space-time continuity comprising the step of increasing the radius (R) of the specific region. The method of claim 5, wherein
The repeating step,

Compensating for the loss of the shoreline under the water; And

Compensating the vegetation index using the space-time continuity further comprising the step of compensating for the loss of the area including the city and the desert on land. The method of claim 1,
The MRT (MODIS Reprojection Tool) is a vegetation index correction method using space-time continuity using the Plate Carree map projection method.

Images