KR20210092866A - Method for utilizing of hyperspectral image using optimal band index - Google Patents

Abstract

The present invention provides a method for utilizing a hyperspectral image using an optimal band index to estimate a real object at high accuracy. According to the present invention, the method comprises the following steps: (a) inputting hyperspectral observation values(Y) of each of a plurality of predetermined points to a hyperspectral observation value input module (310); (b) inputting a photographed image having hyperspectral information to a photographed image input module (210) and checking a hyperspectral photographing value (I(x), which is a pixel value for a plurality of wavelengths (x) at each of the plurality of predetermined points ); (c) selecting, by a band index calculation module (320), a pair of any two wavelengths (λ_1, λ_2) from a plurality of wavelengths (x) included in the hyperspectral imaging value and checking a hyperspectral imaging value (I(λ_1), I(λ_2)) for each of the pair of wavelengths; (d) calculating, by a band index calculation module (320), a band index (X) by dividing a difference between the two identified hyperspectral imaging values (I(λ_1), I(λ_2)) of the two identified hyperspectral imaging values I(λ_1) by the sum of the two identified hyperspectral imaging values (I(λ_1), I(λ_2)); and (e) assuming, by a regression analysis module (330), that the hyperspectral observation value (Y) and the calculated band index (X) have a linear relationship, and using the hyperspectral observation value (Y)at the plurality of points and the calculated band index (X) to calculate a and b of a regression equation (Y = aX + b), thereby generating the regression equation (Y = aX + b).

Description

Method for utilizing of hyperspectral image using optimal band index}

The present invention relates to a hyperspectral image utilization method using a band index, and specifically, by using an observation value and a hyperspectral image to check a band index that best represents a real object, a point where there is no observed value also relates to a method that can identify objects with high accuracy using only hyperspectral images.

A general image sensor included in a camera detects only information in the visible region, whereas when a hyperspectral sensor is used, it is possible to secure hundreds of times more information including ultraviolet and infrared regions. there is.

In the case of using hyperspectral information including a lot of information, it is possible to precisely identify the object, water depth, environment, etc. of the photographed point only with the corresponding image. To this end, a database that matches what the actual band for each wavelength means is required. For example, if the values of 8 bands (412, 490, 512, 555, 660, 680, 765, 875 nm) are known by the Goestationary Ocean Color Imager (GOGI) launched in 2010 , a database on whether each value means what kind of sea color is already secured. Using this, when hyperspectral information for each point is confirmed by taking hyperspectral images of the ocean, the value at the corresponding wavelength and the sea color matching it are checked from the database, so that the sea color can be determined only with the hyperspectral image without directly observing the sea color. will be estimated

In addition, since it is not possible to verify and database the observation values for each wavelength at all points, a regression equation is used. That is, if a regression equation is constructed using a large number of data, the actual value can be estimated using the band for each wavelength identified in the hyperspectral image by using the regression equation even at a location that is not in a database.

However, it is known that a method of estimating information using a photographing value of a specific wavelength (ie, a band for each wavelength) in this way has a certain level of error. As the causes of such errors, errors occurring in the hyperspectral information capturing step (for example, a situation in which the captured value does not represent the actual location value due to the influence of reflected light), and errors occurring in the image correction step (e.g., Errors in geometric correction, errors in image registration, etc.), errors in the database (eg, when the photographing value is correct but identification information is incorrect), and other external factors are suggested.

Moreover, when a value is estimated through a regression equation constructed using a low-accuracy value, this error gradually increases.

Accordingly, there is an increasing demand for a new identification method capable of increasing accuracy by reducing errors.

Review the relevant patent literature.

Korean Patent Registration No. 10-1463351 discloses a technique for increasing the speed of information calculation by detecting a singular region using a band ratio in a hyperspectral image and excluding the detected region in a hyperspectral image. Only the point of using the band ratio and the present invention to be described below is common, but the overall purpose, configuration, and effect of the invention are different.

Korean Patent Registration No. 10-1806488 discloses a technology for identifying a dark body by mounting a hyperspectral image sensor on a drone. Korean Patent No. 10-1744662 discloses a technology for classifying plants using hyperspectral images. To this end, hyperspectral image standard data for various dark bodies or plants must be secured in advance, and the photographed object is identified by comparing the photographed image thereto. However, it does not solve the fundamental problems identified during the shooting stage, such as a problem of decreasing accuracy due to the shaking of the drone, and a problem of increasing the efficiency of image matching.

US Patent No. 7,181,055 proposes an algorithm for registering hyperspectral images. Here, a reflectance image is used to increase the registration accuracy.

Korean Patent Registration No. 10-1463351 Korean Patent No. 10-1806488 Korean Patent No. 10-1744662 US Patent No. 7,181,055

The present invention has been devised to solve the above problems.

Specifically, we propose a novel hyperspectral information identification method to obtain accurate hyperspectral information and prevent a situation in which the database for the corresponding photographing value does not reflect accurate identification information even when the photographing value of a specific wavelength is accurate. want to

In particular, it is intended to propose a novel hyperspectral image utilization method using a band index, which is a photographing value of a pair of two wavelengths, rather than using a photographing value of a specific wavelength.

In one embodiment of the present invention for solving the above problems, (a) the hyperspectral observation value input module 310, each of a plurality of predetermined points hyperspectral observation value (Y) is input; (b) a photographed image having hyperspectral information is input to the photographed image input module 210, and a hyperspectral photographing value (I(x), which is a pixel value for a plurality of wavelengths (x) at each of the plurality of predetermined points ) is confirmed; _{(c) the band index calculation module 320 selects a pair of any two wavelengths (λ 1} , λ ₂ ) among a plurality of wavelengths (x) included in the hyperspectral photographing value, and hyperspectral for each checking the photographing values (I(λ ₁ ), I(λ ₂ )); (d) the band index calculation module 320 calculates the difference between the two identified hyperspectral imaging values (I(λ ₁ ), I(λ ₂ )) of the identified two hyperspectral imaging values (I ( λ ₁ ), dividing by the sum of I (λ ₂ )) calculating a band index (band index) (X); and (e) the regression analysis module 330 assumes that the hyperspectral observation value (Y) and the calculated band index (X) have a linear relationship, and the hyperspectral observation value ( Using Y) and the calculated band index (X), by calculating a and b of the regression equation (Y = aX + b), the band comprising the step of generating a regression equation (Y = aX + b) A method of using hyperspectral images using an index is provided.

In addition, after step (e), (f) the regression analysis module 330 selects any other than the pair of two wavelengths in step (c) among the plurality of wavelengths included in the hyperspectral imaging value. generating a regression equation for each of a plurality of pairs of two wavelengths by selecting a wavelength of , checking a hyperspectral imaging value thereof, and repeating steps (d) to (e); (g) selecting, by the regression equation determination module 340, a regression equation having the highest correlation coefficient among the generated plurality of regression equations; and (h) determining, by the regression expression determination module 340, two wavelengths applied to the band index (X) used in the regression expression selected in step (g) as a pair of optimal band index wavelengths. It is preferable to do

In addition, after step (h), (i) the observation value estimation module 350 corresponds to the pair of optimal band index wavelengths determined in step (h) in hyperspectral imaging values of points other than the plurality of points. calculating the band index of the corresponding point by checking the two hyperspectral imaging values and dividing the difference between the imaging values by the sum of the imaging values; and (j) estimating, by the observation value estimation module 350, the observed value of the corresponding point by inputting the calculated band index into the regression equation selected in the step (g).

In addition, in the step (g), the regression equation determination module 340 calculates the correlation coefficients for the generated plurality of regression equations, and converts each wavelength in the pair of two wavelengths to the X-axis and the Y-axis. and generating a correlation coefficient map to: (h), wherein the regression equation determination module 340 selects a point having the highest correlation coefficient using the correlation coefficient map generated in the step (g). It is preferable to include the step of determining and determining each wavelength constituting the corresponding point as a pair of the optimal band index wavelength.

In addition, the step (a) may include: (a1) setting, by the flight path setting module 220, the flight path of the unmanned aerial vehicle using the collection area; and (a2) obtaining the hyperspectral information for each preset image unit through line scanning using a hyperspectral sensor while the unmanned aerial vehicle is flying along the set flight path.

Another embodiment of the present invention for solving the above problems is stored in a computer-readable recording medium, and provides a program for executing the above-described method, and a computer-readable recording medium in which the code is recorded.

By the method according to the present invention, by utilizing the band index, it is possible to estimate a real object with high accuracy through the captured value in the hyperspectral information.

Conventionally, only one wavelength having a value identified as closest to the wavelength of the real object is used, so if the matching is inaccurate or there is an error in the measured value, an error in estimating an object different from the real object may occur. There was a downside.

The present invention utilizes two wavelengths with high correlation, but when the correlation is large, considers that there is a positive correlation but also a negative correlation, and utilizes two bands to estimate the value of a real object through a matching process. , which can further increase the accuracy.

This method is computer-programmed, and when observation values of a plurality of points and captured hyperspectral images are input, a regression equation can be automatically determined. It has the advantage of being able to identify.

1 shows a system for carrying out a method according to the invention;
2 is a flowchart illustrating a method according to the present invention.
3 is a view for explaining the band index of the method according to the present invention.
4 is a diagram for explaining a process in which regression analysis is performed using a band index and an observation value in the method according to the present invention.
5 shows a correlation map established by the method according to the present invention.
6 illustrates a method for selecting a regression equation using a correlation map by the method according to the present invention.
7 shows a screen of a program in which the method according to the present invention is implemented.

Hereinafter, the present invention will be described in more detail with reference to the drawings.

Description of the system

A system in which the method according to the present invention is performed may be divided into an image processing unit 200 and an image analysis unit 300 . The image processing unit 200 performs a function of processing a photographed image taken by an unmanned aerial vehicle (not shown) such as a drone, and the image analysis unit 300 uses the pre-processed image here to perform a regression analysis using the band index. , and performs the function of estimating observations.

The image processing unit 200 includes a communication module 210 , a flight path setting module 220 , a captured image input module 230 , a normalization performing module 240 , a geometric correction performing module 250 , an image matching performing module 260 ) and a hyperspectral image output module 270 .

The communication module 210 communicates with the unmanned aerial vehicle (not shown) by wire or wirelessly so that the captured image input module 230 receives the hyperspectral information obtained from the hyperspectral sensor (not shown). The received information may include location information.

The flight path setting module 220 performs a function of setting a flight path of an unmanned aerial vehicle (not shown) to secure the image based on a collection area for which hyperspectral information is to be collected.

The normalization performing module 240 performs a normalization function of the secured hyperspectral information. In all geometric transformations, rotation, magnitude, parallelism preservation, etc. occur with respect to the origin (0, 0), so normalization of each ordered pair of XY coordinates present in each image is necessary.

The geometric correction performing module 250 performs geometric correction for compensating for distortion caused by geometric conditions such as curvature of the ground in the secured hyperspectral information.

The image matching performing module 260 performs a function of matching a plurality of secured hyperspectral information. To this end, an integral image operation module 261 and a corresponding point confirmation module 262 are provided.

The hyperspectral image output module 270 performs a function of outputting the normalized, geometrically corrected, and image-matched hyperspectral image as a final result.

Meanwhile, such an image processing method is only an example, and the hyperspectral image may be processed by any conventionally known processing method.

The image analysis unit 300 includes a hyperspectral observation value input module 310 , a band index calculation module 320 , a regression analysis module 330 , a regression expression determination module 340 , and an observation value estimation module 350 . do.

The hyperspectral observation value input module 310 performs a function of inputting actual hyperspectral observation values for a plurality of predetermined points. That is, the actual observation value for confirming the correlation with the photographed value secured in the hyperspectral image is input.

The band index calculation module 320 calculates the band index through the captured values (eg, pixel values) secured from the hyperspectral image from the image processing unit 200 .

The regression analysis module 330 calculates a regression equation by performing regression analysis of the observed values and the band index.

The regression equation determination module 340 calculates the correlation of a plurality of regression equations, determines the regression equation with the highest correlation, and determines the optimal band index through this.

The observation value estimation module 350 uses the regression equation determined by the regression equation determination module 340 to input a specific hyperspectral image (that is, when an arbitrary band index is input), based on the observation value It estimates and performs the function of identifying a real object based on this.

A method of performing a specific function of each module will be described later.

description of the method

A method of utilizing a hyperspectral image according to the present invention will be described with reference to FIGS. 2 and 4 to 7 .

First, a hyperspectral image must be secured. For example, after the flight path setting module 220 of the image processing unit 200 sets the flight path of the unmanned aerial vehicle (not shown) using the collection area, the unmanned aerial vehicle (not shown) follows the set flight path. It may be a hyperspectral image obtained for each preset image unit through line scanning using a hyperspectral sensor while flying. These hyperspectral images contain hyperspectral information.

In addition, each hyperspectral observation value (Y) should be confirmed for each of a plurality of predetermined points (S210). That is, the hyperspectral observation value Y of each of a plurality of predetermined points is input to the hyperspectral observation value input module 310 . Here, for the accuracy of the regression analysis, it is preferable that the number of points at which the observed value Y is confirmed is greater.

Next, a photographed image having hyperspectral information is input to the photographed image input module 210, and a hyperspectral photographing value (I) that is a pixel value for a plurality of wavelengths (x) at each of a plurality of predetermined points where observation values were input (x)) is confirmed (S220).

_{Next, the band index calculation module 320 selects a pair (λ 1} , λ ₂ ) of any two wavelengths among a plurality of wavelengths (x) included in the hyperspectral imaging value, and a hyperspectral imaging value for each (I(λ ₁ ), I(λ ₂ )) is checked (S230). At this time, the difference between the two identified hyperspectral imaging values (I(λ ₁ ), I(λ ₂ )) is calculated as the difference between the two identified hyperspectral imaging values (I(λ ₁ ), I(λ ₂ )). Divide by the sum to calculate a band index (X) (S240) (refer to FIG. 3). If this is expressed as an equation, it is as Equation 1 below.

In other words, the band index is calculated for each point, and the hyperspectral observation value (Y) and the band index (X) are confirmed, which is performed individually for each of a plurality of points.

Next, the regression analysis module 330 assumes that the hyperspectral observation value (Y) and the calculated band index (X) have a linear relationship, and the hyperspectral observation value (Y) confirmed in step S210 at each of a plurality of points ) and the band index (X) calculated in step S240, a and b of the regression equation (Y = aX + b) are calculated to generate a regression equation (Y = aX + b) (S250). Here, a specific formula for calculating a is the same as Equation 2 below, and when the value of a is calculated, the value of b is naturally calculated. n is the number of observations, X is the band index, and Y is the observation. The upper bar means the average value (this is the same for all equations below).

4 shows an example thereof. Here, the regression equation calculated using the band index (X) and the observed value (Y) at five points (P1, P2, P3, P4, P5) is shown in the lower right corner.

Meanwhile, this process may be performed for each pair while changing the two wavelengths selected in step S230. Even if the predetermined points are the same and the observed values Y at each point are the same, if two wavelengths are selected differently, the final calculated regression equation is different.

Specifically, the regression analysis module 330 selects an arbitrary wavelength other than the first selected pair of two wavelengths, checks the hyperspectral imaging value for this, and repeats steps S240 to S250, a plurality of regression equations to generate (S260).

Next, the regression equation determination module 340 calculates the correlation coefficients of the generated multiple regression equations, and selects a regression equation having the highest correlation coefficient among them ( S270 ). At this time, two wavelengths applied to the band index (X) used in the selected regression equation are determined as a pair of optimal band index wavelengths.

For example, if the number of wavelengths included in the hyperspectral information is N, the number of pairs of two wavelengths is _N C ₂ by the combination, so the regression equation and the correlation coefficient are calculated by the corresponding number, The optimal regression equation and band index and its wavelength will be calculated.

On the other hand, this method may be achieved using a correlation coefficient map. An example of a correlation coefficient map is shown in FIG. 5 . Here, the regression expression determination module 340 calculates the correlation coefficients for the generated plurality of regression equations, but in a pair of two wavelengths, one wavelength is the X-axis and the other wavelength is the Y-axis. The generated correlation coefficient map is shown.

Next, the regression formula determination module 340 determines a point having the highest correlation coefficient using the correlation coefficient map, and determines each wavelength constituting the corresponding point as a pair of the optimal band index wavelengths.

6 shows the concept of automating this scheme. By extracting the imaging value for each wavelength, which is a spectral characteristic, and calculating each band index through this, and regression analysis using the observed values, a correlation coefficient map is generated, and by calculating the correlation coefficient for each pair of wavelengths, the optimal The band index wavelength and optimal regression equation are determined.

7 shows an output screen of a program implementing this. The correlation coefficient map on the right is shown, and it can be seen that the correlation coefficient and regression equation for each wavelength are automatically calculated below it.

Through this process, when the optimal regression equation and the optimal band index are determined, it is possible to estimate the observed value only from the wavelength and the captured value of the hyperspectral information included in the hyperspectral image even at a position where there is no observed value.

For example, when the observation value estimation module 350 calculates the band index by checking the two hyperspectral imaging values corresponding to the pair of optimal band index wavelengths among the photographing values included in the hyperspectral image, the optimal regression By entering into the equation, the observation value of the corresponding point can be estimated with high accuracy.

In the above, the present specification has been described with reference to the embodiments shown in the drawings so that those skilled in the art can easily understand and reproduce the present invention, but these are merely exemplary, and those skilled in the art may make various modifications and equivalent other modifications from the embodiments of the present invention. It will be appreciated that embodiments are possible. Therefore, the protection scope of the present invention should be defined by the claims.

200: image processing unit
210: communication module
220: flight path setting module
230: captured image input module
240: regularization performing module
250: geometric correction performing module
260: image registration performing module
261: integral image operation module
262: correspondence point confirmation module
270: hyperspectral image output module
300: image analysis unit
310: hyperspectral observation value input module
320: band index calculation module
330: regression analysis module
340: regression expression determination module
350: observation value estimation module

Claims (7)

(a) inputting hyperspectral observation values (Y) of each of a plurality of predetermined points to the hyperspectral observation value input module 310;
(b) a photographed image having hyperspectral information is input to the photographed image input module 210, and a hyperspectral photographing value (I(x), which is a pixel value for a plurality of wavelengths (x) at each of the plurality of predetermined points ) is confirmed;
_{(c) the band index calculation module 320 selects a pair of any two wavelengths (λ 1} , λ ₂ ) among a plurality of wavelengths (x) included in the hyperspectral photographing value, and hyperspectral for each checking the photographing values (I(λ ₁ ), I(λ ₂ ));
(d) the band index calculation module 320 calculates the difference between the two identified hyperspectral imaging values (I(λ ₁ ), I(λ ₂ )) of the identified two hyperspectral imaging values (I( λ ₁ ), dividing by the sum of I (λ ₂ )) calculating a band index (band index) (X); and
(e) the regression analysis module 330 assumes that the hyperspectral observation value (Y) and the calculated band index (X) have a linear relationship, and the hyperspectral observation value (Y) at the plurality of points ) and the calculated band index (X), by calculating a and b of the regression equation (Y = aX + b), comprising the step of generating a regression equation (Y = aX + b),
A method of using hyperspectral images using band index.
The method of claim 1,
After step (e),
(f) the regression analysis module 330 selects an arbitrary wavelength other than the pair of two wavelengths in step (c) from among the plurality of wavelengths included in the hyperspectral imaging value, generating a regression equation for each of a plurality of pairs of two wavelengths by verifying a spectroscopic image and repeating steps (d) to (e);
(g) selecting, by the regression equation determination module 340, a regression equation having the highest correlation coefficient among the generated plurality of regression equations; and
(h) the regression formula determination module 340, further comprising the step of determining the two wavelengths applied to the band index (X) used in the regression equation selected in step (g) as a pair of optimal band index wavelengths doing,
A method of using hyperspectral images using band index.
3. The method of claim 2,
After step (h),
(i) the observation value estimation module 350 confirms two hyperspectral imaging values corresponding to the optimal band index wavelength pair determined in step (h) from the hyperspectral imaging values of points other than the plurality of points and calculating the band index of the corresponding point by dividing the difference between the photographing values by the sum of the photographing values; and
(j) the observation value estimation module 350, inputting the calculated band index into the regression equation selected in the step (g), further comprising the step of estimating the observed value of the point,
A method of using hyperspectral images using band index.
4. The method of claim 3,
The step (g) is,
generating, by the regression equation determination module 340, correlation coefficients for the generated plurality of regression equations, and generating a correlation coefficient map in which each wavelength is an X-axis and a Y-axis in a pair of two wavelengths; including,
The step (h) is,
The regression expression determination module 340 determines a point having the highest correlation coefficient using the correlation coefficient map generated in step (g), and sets each wavelength constituting the point to the pair of optimal band index wavelengths. comprising the step of determining
How to determine the optimal band index for hyperspectral images
The method of claim 1,
The step (a) is,
(a1) the flight path setting module 220, setting the flight path of the unmanned aerial vehicle using the collection area; and
(a2) the unmanned aerial vehicle flying along the set flight path, using a hyperspectral sensor, through line scanning, comprising the step of acquiring the hyperspectral information for each preset image unit,
A method for determining the optimal band index of hyperspectral images.
stored on a computer-readable recording medium.
A program for executing a method according to any one of claims 1 to 5.
A computer-readable recording medium in which a program code for executing the method according to any one of claims 1 to 5 is recorded.

Images