KR101958725B1 - METHOD and apparatus for mapping IMAGE color USING MARS - Google Patents

Abstract

Disclosed is a method for color mapping between images. According to the present invention, the method for color mapping between images comprises the steps of: receiving a target image and a reference image; and performing color mapping of the target image based on an algorithm of multivariate adaptive regression splines (MARS) using the target image and the reference image.

Description

METHOD AND APPARATUS FOR IMAGE COLOR USING MARS [0002]

The present invention relates to a method for performing image color mapping using Multivariate Adaptive Regression Splines (MARS).

A Synthetic Aperture Radar (SAR) image is an image that provides information on the surface of the earth using a propagation signal reflected from an observation area. Synthetic aperture radar images are images independent of solar irradiance and weather conditions, unlike optical images. For this reason, studies are underway to improve the usability of synthetic aperture radar images.

The most active research to improve the usability of synthetic aperture radar images is the colorization study to improve the readability of the images. Generally, the colorization of synthetic aperture radar images can be classified into coloring through channel combination of Polarimetric SAR (Polarimetric SAR) and colorization through image fusion.

In the case of colorization through a channel combination of a polarizing system synthetic aperture radar image, a method of performing decomposition of a polarizing system synthetic aperture radar image based on a scattering mechanism and assigning each acquired channel to an RGB band is performed. This method has a limitation in that the readout power is improved when compared with the conventional synthetic aperture radar image, but it can not contain the color of the actual optical image and can not be applied to synthetic aperture radar images other than the polarimeter synthetic aperture radar.

In the case of colorization through image fusion, a synthetic aperture radar image is assumed to be a panchromatic image and fused with a multispectral image. This method is complementary to the information of the synthetic aperture radar image and the multispectral image, and improves the visual reading performance such as feature extraction and object recognition, but there is a limitation that there is a distortion in the fusion result .

At this time, synthetic aperture radar images and multispectral images must be modeled as nonlinear relations in order to model the two image relations due to differences in geometrical, spectral, and sensor characteristics.

BACKGROUND ART [0002] The background of the present invention is disclosed in Japanese Patent Application Laid-Open No. 2013-272474.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a method of performing color mapping between images based on a MARS algorithm.

It is another object of the present invention to provide a method of performing color mapping on a pixel value of a synthetic aperture radar image by establishing a color relationship between a synthetic aperture radar image and a multispectral image.

At this time, synthetic aperture radar images and multispectral images must be modeled as nonlinear relations in order to model the two image relations due to differences in geometrical, spectral, and sensor characteristics. Accordingly, it is an object of the present invention to provide a method of performing color mapping by establishing non-linear color relations using MARS.

It should be understood, however, that the technical scope of the embodiments of the present invention is not limited to the above-described technical problems, and other technical problems may exist.

According to an aspect of the present invention, there is provided a method for performing an inter-image color mapping, the method comprising: receiving a target image and a reference image; And performing color mapping of the target image based on a Multivariate Adaptive Regression Splines algorithm.

According to an embodiment of the present invention, the target image may be a monochrome image, and the reference image may be a color image.

According to one embodiment of the present application, the object image may be a synthetic aperture radar (SAR) image.

According to an embodiment of the present invention, performing the color mapping includes: selecting at least one reference point for color relation establishment from the target image and the reference image; classifying the reference image into a plurality of classes The method includes: establishing a MARS regression equation based on pixel values and characteristic parameter values of the reference points for each of the plurality of classes; and assigning a pixel value and a characteristic variable value of the target image to the MARS regression formula established for each class And obtaining a color-mapped target image.

According to an embodiment of the present invention, the step of selecting the reference point may include: obtaining a gray-scale image of the reference image; and performing image differencing between the gray- To select an invariant pixel (Invariant pixel).

According to one embodiment of the present invention, the invariant point may be a reference point.

According to one embodiment of the present application, the plurality of classes may include water, a city, a desolate area, a forest, a shaded area by forest, and a granite area.

According to an embodiment of the present invention, the step of establishing the MARS regression formula includes obtaining a pixel value and a characteristic variable value of the reference point, and establishing a MARS regression equation for each color band, May include a RED band, a GREEN band, and a BLUE band.

According to an embodiment of the present invention, there is provided an apparatus for performing an inter-image color mapping, the apparatus comprising: an image receiving unit for receiving a target image and a reference image; And a color mapping unit for performing color mapping of the target image.

According to an embodiment of the present invention, the color mapping unit includes a reference point selecting unit for selecting at least one reference point for establishing color relation from the target image and the reference image, a classifying unit for classifying the reference image into a plurality of classes, A regression equation establishing unit that establishes a MARS regression equation based on the pixel values of the reference points and the characteristic variable values for each of the plurality of classes, And a color image acquiring unit for acquiring a color-mapped target image by substituting pixel values and characteristic parameter values of the target image into the MARS regression equation.

Further, it is possible to provide a computer-readable recording medium on which a program for causing a computer to execute the above-described method is recorded.

The above-described task solution is merely exemplary and should not be construed as limiting the present disclosure. In addition to the exemplary embodiments described above, there may be additional embodiments in the drawings and the detailed description of the invention.

According to the above-described task resolution means, it is possible to improve the reading power of a target image by providing a method of performing an inter-image color mapping based on the MARS algorithm.

According to the above-mentioned problem solving means of the present invention, it is possible to establish a non-linear color relation between images based on the MARS algorithm.

In addition, the present invention provides a method of performing color mapping on a pixel value of a synthetic aperture radar image by establishing a color relationship between a synthetic aperture radar image and a multispectral image, thereby improving the readability of a synthetic aperture radar (SAR) image .

FIG. 1 is a block diagram schematically illustrating a color mapping performance system including an apparatus for performing an inter-image color mapping according to an exemplary embodiment of the present invention. Referring to FIG.
2 is a block diagram illustrating an apparatus for performing an inter-image color mapping according to an exemplary embodiment of the present invention.
3 is a block diagram illustrating a color mapping unit of an apparatus for performing an inter-image color mapping according to an exemplary embodiment of the present invention.
FIG. 4 is a diagram illustrating an example of a color-mapped target image produced by a method of performing an inter-image color mapping according to an exemplary embodiment of the present invention.
FIG. 5 is an exemplary view of an invariant point selection image selected by a method of performing inter-image color mapping according to an exemplary embodiment of the present invention.
FIG. 6 is an exemplary view illustrating an image classified into classes according to an image-to-image color mapping method according to an exemplary embodiment of the present invention.
FIG. 7 is an exemplary diagram illustrating a method of evaluating a color-mapped target image, which is produced by the method of performing an inter-image color mapping according to an embodiment of the present invention.
8 is a first flowchart illustrating a flow of a method of performing an inter-image color mapping according to an embodiment of the present invention.
9 is a second flowchart illustrating a flow of a method of performing an inter-image color mapping according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention. It should be understood, however, that the present invention may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In the drawings, the same reference numbers are used throughout the specification to refer to the same or like parts.

Throughout this specification, when a part is referred to as being "connected" to another part, it is not limited to a case where it is "directly connected" but also includes the case where it is "electrically connected" do.

It will be appreciated that throughout the specification it will be understood that when a member is located on another member "top", "top", "under", "bottom" But also the case where there is another member between the two members as well as the case where they are in contact with each other.

Throughout this specification, when an element is referred to as "including " an element, it is understood that the element may include other elements as well, without departing from the other elements unless specifically stated otherwise.

1 is a block diagram schematically illustrating a color mapping performance system including an apparatus for performing an inter-image color mapping according to an embodiment of the present invention (hereinafter, referred to as an 'apparatus for performing an inter-image color mapping') 1000, 2 is a block diagram illustrating an apparatus 1000 for performing an inter-image color mapping according to an exemplary embodiment of the present invention. 8 is a first flowchart illustrating a flow of a method of performing an inter-image color mapping according to an embodiment of the present invention.

Referring to FIGS. 1, 2, and 8, a method of performing inter-image color mapping according to an exemplary embodiment of the present invention includes receiving a target image 110 and a reference image 120 (S100) (S200) of performing a color mapping of the target image 110 based on a Multivariate Adaptive Regression Splines (MARS) algorithm using the reference image 110 and the reference image 120.

In step S100, the target image 110 is an image to be subjected to color mapping. The target image 110 may be a monochrome image. In addition, the target image 110 may include a relatively low pixel value as compared with the reference image 120, or a video image having relatively low readout power due to distortion due to lens conditions, weather conditions, . Illustratively, the subject image 110 may be an image obtained in a Synthetic Aperture Radar (SAR) 900, but is not limited thereto.

The synthetic aperture radar 900 is a radar that observes the ground and the ocean in the air. Specifically, the synthetic aperture radar 900 is a radar system that generates a terrestrial topographical map by sequentially radiating a radio wave signal with respect to the ground and the ocean, and then synthesizing a time difference in which the radio wave signal is reflected from the observation area and returned on a first- In addition, the synthetic aperture radar image is an image created through the radar system. The synthetic aperture radar image can be made through the radar system not only in normal mode, but also in daylight and in bad weather.

In step S100, the reference image 120 is an image to be referred to when color mapping is performed on the target image 110. [ The reference image 120 may be a color image as a multispectral image. Also, the reference image 120 may be an image having a relatively high pixel value or providing a relatively high reading power as compared with the target image 110. Illustratively, the reference image 120 may be an image acquired from a satellite or an aircraft, but is not limited thereto.

For a detailed description of the present application, specific examples of a target image 110 and a reference image 120 are presented. 4, the target image 110 shown in FIG. 4 is a Radarsat-1 image in the Daejeon metropolitan area of Korea, and the reference image 120 shown in FIG. 4 is a Landsat-5 TM image in the Daejeon metropolitan area of Korea . The Radarsat-1 image is an image of the earth's surface using a synthetic aperture radar, and the Landsat-5 TM image is an image of the earth's surface using an earth orbiting satellite.

1 and 2, the image receiving unit 100 of the inter-image color mapping performing apparatus 1000 receives the target image 110 and the reference image 120 through the network 500 have. Illustratively, the image receiving unit 100 can receive an image generated through the synthetic aperture radar 900 through the network 500. [ In addition, the image receiving unit 100 can receive the composite aperture radar image stored in the object and reference image storage unit 850 in advance through the network 500. Meanwhile, the image receiving unit 100 can receive an image created through a satellite or an aircraft through the network 500, and the image receiving unit 100 is stored in the object and reference image storing unit 850 in advance The satellite image or the aerial image can be received through the network 500. [

In step S200, the color mapping unit 200 of the inter-image color mapping apparatus 1000 performs color mapping of the target image 110 based on the MARS algorithm using the target image 110 and the reference image 120 Can be performed. The MARS algorithm is a form of regression analysis, a non-parametric regression analysis algorithm. In addition, the MARS algorithm is a nonlinear relationship modeling algorithm that can automatically model nonlinearities and interactions between variables. MARS establishes a regression equation by building relationships in a series of coefficients and primitive functions, instead of assumptions about the functional relationships between dependent and independent variables. The basic principle is a recursive partitioning and multi-stage regression-based algorithm that uses the spline function to sort the data by dividing the input region and establishing and using self regression equations for each region. MARS is a simple and highly predictive method for high dimensional data relationships and nonlinear relationships, and has the advantage of being easy to interpret the generated model, so prediction and simulation can be utilized. The MARS algorithm and the MARS regression equation are described in detail below.

FIG. 3 is a block diagram illustrating a color mapping unit 200 of an apparatus 1000 for performing an inter-image color mapping according to an exemplary embodiment of the present invention. FIG. 2 is a second flowchart showing the flow.

3 and 9, the step of performing color mapping (S200) includes a step S210 of selecting at least one reference point for color relation establishment from the target image 110 and the reference image 120, A step S220 of classifying the image 120 by a plurality of classes 600, a step S230 of establishing a MARS regression equation based on pixel values and characteristic variable values of the reference point 700 for each of the plurality of classes 600 And a step S240 of obtaining the color-mapped target image 950 by substituting the pixel value and the characteristic parameter value of the target image 110 into the MARS regression equation established for each of the classes 600 and 600.

According to an embodiment of the present invention, the step S210 of selecting a reference point may include a step of acquiring a gray-scale image 130 of the reference image 120 and a step of acquiring a gray- 110), and selecting an invariant pixel as a reference point by performing image differencing.

The reference point may be a pixel serving as a reference for establishing a color relation between the target image 110 and the reference image 120. [

FIG. 5 is an exemplary view of an invariant point selection image 140 selected by a method of performing inter-image color mapping according to an exemplary embodiment of the present invention.

The gray scale image 130 can be obtained in the reference image 120 as a single sample image in which the value of each pixel represents only the amount of light (amount of light). According to one embodiment of the present invention, the color mapping unit 200 may acquire a gray scale image 130 of the reference image 120 using a luminance method. The method of acquiring the gray scale image 130 of the reference image 120 may further include a colorimetric conversion method or a Luma coding method. Illustratively, referring to FIG. 5, the gray scale image 130 shown in FIG. 5 is an image obtained using the above-described photometric method for the Landsat-5 TM image.

In step S214, the color mapping unit 200 may perform an image balance method between the gray-scale image 130 of the reference image 120 and the target image to select an invariant point. Image Batching is an image processing technique used to determine the change between images. The difference between the images can be calculated based on the difference between the pixels of each image. In order to utilize the image balancing method, first, the luminance values of images to be subjected to the image balancing method must be made compatible.

According to one embodiment of the present invention, an invariant point between the gray scale image 130 and the target image 110 can be selected according to the image balance method. The invariant point means a pixel determined to have not changed according to the image balance method between the gray scale image 130 and the target image 110. Meanwhile, the color mapping unit 200 can select an invariant point as a reference point. 5, an invariant point between the Radarsat-1 image and the Landsat-5 image is selected according to the image balance method, This is the image shown.

FIG. 6 is an exemplary diagram of an image classified by class 600 according to an image-to-image color mapping method according to an embodiment of the present invention.

In step S220, the color mapping unit 200 may classify the reference image 120 according to a plurality of classes 600. For example, The class 600 may be a collection of pixels having a common property in an image. For example, the class 600 may be a legend that can classify pixels having the common attributes, such as mountainous areas, water areas, urban areas, and desert areas. According to one embodiment of the present invention, the color mapping unit 200 can determine the class and the number of the classes according to the color resolution of the reference image 120. As the resolution of the reference image 120 increases, the number of classifiable classes may increase. As the resolution of the reference image 120 increases, the number of classifiable classes increases. As described later, a pixel value and a characteristic variable value corresponding to a reference point are obtained for each class and a MARS regression equation is established to establish a color relationship , And further improved coloring is possible.

According to an embodiment of the present invention, the method of classifying the reference image 120 by a plurality of classes 600 may be a maximum likelihood method. Since the maximum method is a method obvious to a person skilled in the art, a detailed description will be omitted.

In addition, according to one embodiment of the present disclosure, the plurality of classes 600 may include water, Urban, Barren, Forest, Shadow caused by forest, And a grain region Crop. The class classification criteria described above are in accordance with USGS-Level II, one of the standards set by the United States Geological Survey (USGS). In addition, the classification of the class may be classified into a plurality of classes based on international standard including the USGS-Level II, but is not limited thereto. According to one embodiment of the present invention, the color mapping unit 200 may classify the reference points and classify the classes so that classes are allocated to each reference point.

Illustratively, referring to FIG. 6, the image shown in FIG. 6 is an image 960 obtained by classifying the Landsat-5 TM image by a plurality of classes 600. FIG. In the image (960) obtained by classifying by class, it can be confirmed that the legend is classified by blue, water area is red, and desolate area is yellow.

According to one embodiment of the present invention, the step of establishing the MARS regression equation (S230) may include obtaining pixel values and characteristic parameter values of the reference point for each class and establishing a MARS regression equation for each color band . The color bands may include RED, GREEN, and BLUE bands.

Specifically, the color mapping unit 200 can obtain the pixel value of the reference point and the characteristic variable value. According to an embodiment of the present invention, the color mapping unit 200 may analyze a target image 110 and a reference image 120 using a known image analysis algorithm to obtain a pixel value and a characteristic variable value. For example, the value of the characteristic variable may be calculated using the contrast coincidence matrix (D) obtained through Contrast, Dissimilarity, Homogeneity, Correlation, Angular Second Moment, Energy, and Entropy GLCM: Gray Level Co-Occurrence Matrix) value. In addition, the characteristic variable value may include Mean Pixel Values and Variance Values, but is not limited thereto. Illustratively, the characteristic variable values can be selected as shown in Table 1 below.

Property Variable Value Induced variable of characteristic variable value The pixel of the target image 110 The pixel value of the target image 110 Concurrency matrix The homogeneity of the pixel neighbors of the 5x5 pixel neighborhood calculated from the target image 110 Concurrency matrix The energy of the (Pixel Neihborhood) adjacent to the 5x5 pixel calculated from the target image 110 Concurrency matrix The ASM (Pixel Neighborhood) of the 5x5 pixel adjacent region calculated from the target image 110 Concurrency matrix The contrast of the 5x5 pixel neighborhood (Pixel Neihborhood) calculated from the target image 110 Concurrency matrix (Pixel neighbors) of the 5x5 pixel adjacent region calculated from the target image 110 Concurrency matrix The entropy of the Pixel Neighborhood of the 5x5 pixel neighborhood calculated from the target image 110 Average pixel value The average value of the pixel neighbors of the 5x5 pixel adjacent region calculated from the target image 110 Change value The change value (Pixel Neihborhood) of the 5x5 pixel adjacent region calculated from the target image 110

The color mapping unit 200 can establish a MARS regression equation based on the pixel values of the reference points and the characteristic variable values for each of a plurality of classes. Since the complexity of the calculation is high and the processing time may be considerably long when the MARS regression equation is established for the entire image, the color mapping unit 200 of the present invention can calculate the pixel values of the reference points and the characteristic variable values By establishing the MARS regression formula on the basis of this, it is possible to lower the computational complexity and improve the processing speed. According to one embodiment of the present invention, the obtained pixel value of the invariant point and the characteristic variable value are substituted into the MARS algorithm, and the MARS regression equation can be established for each color band. The color bands may include RED bands, GREEN bands, and BLUE bands as described above. In other words, three color relation regression equations can be established for each color band. In this case, the target image 110 and the reference image 120 can correspond to 1: 3 colors RED, GREEN, and BLUE.

Here, the MARS algorithm and the MARS regression expression will be described in detail. The MARS algorithm is a nonlinear relationship modeling algorithm that can automatically model nonlinearities and interactions between variables, as described above. The basic principle of the MARS algorithm is to divide the input region and establish a self regression equation for each region. It uses a spline function, recursive partitioning, and multi- Stage Regression). Meanwhile, the spline function is a mathematical tool for creating a smooth curve using a given plurality of control points (Knots).

The MARS algorithm can be expressed as Equation (1) below.

는 기본 함수(Basic Function),

및

는 추정 계수들이고,

은

의 시리즈 넘버(Series Number)이다. 또한,

의 계수

은 하기의 수학식 2 및 수학식 3으로 표현될 수 있다.In Equation (1)

Is a basic function,

And

Are the estimation coefficients,

silver

(Series Number). Also,

Coefficient of

Can be expressed by the following equations (2) and (3).

는 상기 제어점(Knot)의 위치이고,

는

번째의 설명 변수이다. 여기서, 상기 제어점(Knot)은 한 데이터 영역의 끝과 다른 영역의 시작을 표시할 수 있다. 상기 제어점의 수, 및 위치는 상기 스플라인 함수에 대해 고정될 수 있으며, MARS 알고리즘(300)에서, 상기 제어점은 순방향(Forward) 및 역방향(Backward)에서의 단계적 검색에 기초하여 결정될 수 있다.In Equations (1) and (2)

Is the position of the control point Knot,

The

. Here, the control point Knot may indicate the beginning of a region different from the end of one data region. The number and position of the control points may be fixed for the spline function and in the MARS algorithm 300 the control points may be determined based on a stepwise search in the Forward and Backward directions.

First, the MARS algorithm can generate a model with an excessive number of control points. Then, a control point having a small contribution to the suitability of the spline function can be eliminated.

와 목표 변수

사이의 관계를 나타내는 함수의 집합으로 표현될 수 있다.The basic function of the MARS algorithm 300 can be utilized to search for control points. Illustratively, the basic function is expressed by the following equation (4)

And target variable

Can be expressed as a set of functions representing the relationship between the two.

및

는 추정 계수들이고,

는 다른 기본 함수인

의 가중치를 고려한 합이다. 아울러, MARS 알고리즘은,

와

형태의 선형 기본 함수(Linear Basic Function)이며,

는 제어점의 위치일 수 있다.In the function of equation (4)

And

Are the estimation coefficients,

Is another basic function

Which is a sum of weights. In addition,

Wow

Linear Basic Function,

May be the position of the control point.

The procedure of the MARS algorithm can be divided into three steps. First, (a) the forward MARS algorithm selects all possible basic functions and corresponding control points, (b) the reverse MARS algorithm removes the basic functions to generate the best combination of existing control points, and finally (c) A smoothing operation may be performed to obtain continuous partition borders.

을 결정하여 집합 C의 모든 함수가 후보가 되도록 할 수 있다. 이 때, MARS 알고리즘에 기 설정한 최대 개수의 조건까지 새로운 쌍의 함수들이 각 단계에서 고려될 수 있다.The step (a) includes, in the initial set, selecting all of the possible basic functions,

So that all the functions of the set C are candidates. At this time, a new pair of functions up to the maximum number of conditions preset for the MARS algorithm can be considered at each step.

In step (b), the reverse MARS algorithm may be performed by removing the basic functions contributing to the minimum residual error. The reverse MARS algorithm can reduce the complexity of the modeling by increasing generalisability. Here, the backward MARS algorithm can be performed through Generalized Cross-Validation (GCV) as shown in Equation (5).

In the step (c), the planarization operation can ensure continuity by eliminating discontinuity in the region boundary.

Based on the principle of the MARS algorithm described above, a method of performing an inter-image color mapping according to an embodiment of the present invention includes calculating a characteristic variable value corresponding to an invariant point selected as a reference point, By assigning a value to the MARS algorithm, a MARS regression equation can be established for each color band.

In step S240, the color mapping unit 200 may acquire the color-mapped target image 950 by substituting the pixel values and the characteristic variable values of the target image 110 into the established MARS regression equation. 4, by applying the pixel values and the characteristic variable values corresponding to all the pixels of the Radarsat-1 image to the established MARS regression equation, the entire color relationship is obtained and the color-mapped Radarsat- 1 image can be obtained. The color-mapped Radarsat-1 image 950 is shown at the far right in FIG. In addition, the color-mapped Radarsat-1 image 950 can significantly improve the reading power as compared to the existing Radarsat-1 image 110, and the spectral characteristics of the Landsat- It can be confirmed that it has a spatial characteristic.

Illustratively, the method of performing the inter-image color mapping can perform a quantitative performance evaluation. In order to perform the quantitative performance evaluation of the method of performing the inter-image color mapping, in the case of spectral quality, a universal image quality index (UIQI), a correlation coefficient (CC) Root Mean Square Error (RMSE) can be utilized, and entropy can be utilized for spatial quality. Illustratively, the quality of the images shown in FIG. 4 may be as shown in Table 2 below.

video Color band UIQI CC RMSE Entropy Radarsat-1 video
(Raw) - 0.0523 0.0817 43.9959 7.0287 Landsat-5 TM Imaging RED
GREEN
BLUE 0.7332
0.6548
0.5501 0.8116
0.7512
0.6290 9.0115
9.0503
9.3048 8.1002
8.1006
8.1039 Color-mapped Radarsat-1 video RED
GREEN
BLUE 0.8213
0.8233
0.7965 0.8334
0.8238
0.7976 8.9746
8.8986
9.0430 8.0598
8.0641
8.0686

1, the color-mapped Radarsat-1 image 950 is compared with the Radarsat-1 image, which is the target image 110, and the generalized image quality index (UIQI ), The correlation coefficient (CC), the mean square root deviation (RMSE), and the entropy were improved by 59.37%, 64.89%, 79.26%, and 13.24%, respectively.

In addition, the color-mapped Radarsat-1 image 950 has a generalized image quality index (UIQI), a correlation coefficient (CC), and an image quality index The mean square root deviation (RMSE) is 16.77%, 8.76% and 1.64% lower, respectively, but the entropy is 0.47% higher.

As a result, it can be seen that the color-mapped Radarsat-1 image 950 significantly improves performance compared to the Radarsat-1 image 110. [ Also, when compared with the evaluation reference image 970, the spectral quality is somewhat lowered, but the spatial quality is improved. This ensures that the color mapped Radarsat-1 image 950 can have a performance substantially similar to the evaluation reference image 970 and also the usefulness of performing color mapping using the MARS algorithm 300 .

Hereinafter, an apparatus 1000 for performing an inter-image color mapping according to an embodiment of the present invention will be described. However, since the apparatus 1000 for performing an image-to-image color mapping is an apparatus for performing the method of performing color mapping according to an embodiment of the present invention, which is the same as the method described above or a technical feature corresponding thereto, The same reference numerals are used for the same or similar components as those of the first embodiment, and redundant explanations will be simplified or omitted.

The apparatus 1000 for performing an inter-image color mapping according to an exemplary embodiment of the present invention may include an image receiving unit 100, and a color mapping unit 200. Illustratively, the inter-image color mapping device 1000 may be, but is not limited to, a computer, a server, or a device equipped with software for performing color mapping between images.

The image receiving unit 100 receives the target image 110 and the reference image 120. The image receiving unit 100 can receive the target image 110 and the reference image 120 through the network 500 as described above. Illustratively, the image receiving unit 100 can receive an image generated through the synthetic aperture radar 900 through the network 500. [ In addition, according to one embodiment of the present invention, the image receiving unit 100 may include a storage unit, and the object image 110 and the reference image 110 may be received through the storage unit.

The network 500 between the image receiving unit 100 and the synthetic aperture radar 900 or between the image receiving unit 100 and the object and reference image storage unit 850 may be provided as a wired and wireless network, Examples include RFID, a 3rd Generation Partnership Project (3GPP) network, a Long Term Evolution (LTE) network, a 5G network, a World Interoperability for Microwave Access (WIMAX) network, the Internet, a LAN (Local Area Network) (LAN), a Wide Area Network (WAN), a Personal Area Network (PAN), a Bluetooth network, a Wi-fi network, a satellite broadcasting network, an analog broadcasting network, a DMB (Digital Multimedia Broadcasting) But is not limited thereto.

The color mapping unit 200 performs a color mapping of the target image 110 based on a Multivariate Adaptive Regression Splines (MARS) algorithm 300 using the target image 110 and the reference image 120.

FIG. 3 is a block diagram illustrating a color mapping unit 200 of an inter-image color mapping device 1000 according to an embodiment of the present invention.

3, the color mapping unit 200 includes a reference point selecting unit 210, a classifying unit 220, a variable obtaining unit 230, a regression formula setting unit 240, and a color image obtaining unit 250, . ≪ / RTI >

The reference point selection unit 210 selects at least one reference point for color relation establishment from the target image 110 and the reference image 120.

The classifying unit 220 classifies the reference image 120 into a plurality of classes.

The variable obtaining unit 230 obtains pixel values and characteristic variable values of reference points and pixels of the image.

The regression equation establishing unit 240 constructs the MARS regression equation based on the pixel values of the reference points and the characteristic variable values for each of the plurality of classes 600.

The color image acquisition unit 250 acquires the color-mapped target image 950 by substituting the pixel value and the characteristic parameter value of the target image 110 into the MARS regression formula established for each class 600. [

The method of performing inter-image color mapping according to one embodiment of the present invention may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions recorded on the medium may be those specially designed and constructed for the present invention or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the present invention, and vice versa.

In addition, the above-described method of performing the inter-image color mapping according to an embodiment of the present invention may be implemented in the form of a computer program or an application executed by a computer stored in a recording medium.

It will be understood by those of ordinary skill in the art that the foregoing description of the embodiments is for illustrative purposes and that those skilled in the art can easily modify the invention without departing from the spirit or essential characteristics thereof. It is therefore to be understood that the above-described embodiments are illustrative in all aspects and not restrictive. For example, each component described as a single entity may be distributed and implemented, and components described as being distributed may also be implemented in a combined form.

The scope of the present invention is defined by the appended claims rather than the detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalents should be construed as being included within the scope of the present invention.

100:
110: target image
120: reference image
130: Grayscale video
200: Color mapping unit
210: Reference point selection section
220:
230: variable acquisition unit
240: regression equation forming section
250: Color image acquiring unit
500: Network
900: Synthetic aperture radar
1000: Color mapping device

Claims (13)

Receiving a target image and a reference image; And
Performing color mapping of the target image based on a Multivariate Adaptive Regression Splines (MARS) algorithm using the target image and the reference image,
Wherein the performing the color mapping comprises:
Selecting at least one reference point for color relation establishment from the target image and the reference image;
Classifying the reference image into a plurality of classes;
Establishing a MARS regression equation based on pixel values and characteristic variable values of the reference points for each of the plurality of classes; And
Acquiring a color-mapped target image by substituting pixel values and characteristic parameter values of the target image into a MARS regression equation established for each class;
To-color mapping. The method according to claim 1,
Wherein the target image is a monochrome image and the reference image is a color image. 3. The method of claim 2,
Wherein the object image is a synthetic aperture radar (SAR) image. delete The method according to claim 1,
Wherein the step of selecting the reference point comprises:
Obtaining a gray-scale image of the reference image; And
Selecting an invariant pixel by performing image differencing between the gray-scale image and the target image,
Image color mapping. 6. The method of claim 5,
Wherein the invariant point is a reference point. The method according to claim 1,
Wherein the plurality of classes comprises water, a city, a desolate area, a forest, a shaded area by forest, and a granite area. The method according to claim 1,
The step of establishing the MARS regression equation comprises:
Obtaining a pixel value and a characteristic variable value of the reference point; And
Establishing a MARS regression equation for each color band,
Lt; / RTI >
Wherein the color band comprises a RED band, a GREEN band, and a BLUE band. An image receiving unit for receiving a target image and a reference image; And
And a color mapping unit for performing color mapping of the target image based on a Multivariate Adaptive Regression Splines (MARS) algorithm using the target image and the reference image,
The color mapping unit includes:
A reference point selecting unit for selecting at least one reference point for establishing a color relation from the target image and the reference image;
A classifying unit for classifying the reference image into a plurality of classes;
A variable obtaining unit obtaining a pixel value and a characteristic variable value of the reference point;
A regression equation establishing unit for establishing a MARS regression equation based on pixel values and characteristic variable values of the reference points for each of the plurality of classes; And
A color image acquiring unit for acquiring a color-mapped target image by substituting pixel values and characteristic parameter values of the target image into a MARS regression equation established for each class;
And an image-to-image color mapping unit. delete 10. The method of claim 9,
Wherein the target image is a monochrome image and the reference image is a color image. 10. The method of claim 9,
Wherein the object image is a synthetic aperture radar (SAR) image. A computer-readable recording medium storing a program for causing a computer to execute the method according to any one of claims 1 to 3 and 5 to 8.

Images