KR100546646B1 - Image distortion compensation method and apparatus - Google Patents

Abstract

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for correcting color distortion and color convergence of a screen caused by optical or mechanical deformation in a display device such as a projection TV, a projector, a monitor, and the like. By dividing into the problem of distortion, and processing each independently, using the minimum order polynomial required for each, greatly reducing the cost of implementation, resulting from mechanical or optical deformation in display devices such as projection TVs, projectors, monitors, etc. Geometric distortions and color convergence distortions can be corrected.

Image Warping, Geometric Distortion, Convergence Distortion

Description

Image distortion compensation method and apparatus

1 is a view showing the structure and optical distortion of the CRT projection TV

2 is a block diagram illustrating an optical distortion correction apparatus using a conventional digital image processing technique

3A to 3H illustrate types of distortions existing in a general display device.

4 is a flowchart of an image warping method according to the present invention.

5 is a general diagram illustrating an example of pairs of corresponding coordinates of control points of a source image and a target image;

6 is a block diagram of a screen distortion correction apparatus according to the present invention;

Explanation of symbols for main parts of the drawings

601: first warping coordinate calculation unit 602: R warping unit

602a: R warping coordinate calculation unit 602b: R mapping unit

603: G warping part 603a: G warping coordinate calculation part

603b: G mapping part 604: B warping part

604a: B warping coordinate calculation unit 604b: B mapping unit

605: projection

The present invention relates to a method and apparatus for correcting geometric distortion and color convergence of a screen by optical or mechanical deformation in a display device such as a projection TV, a projector, a monitor, and the like.

As demand for large-screen TVs surges, projection TVs are diversifying into CRT, LCD, and DLP products.

Figure 1 shows an example of the structure of the CRT projection TV, the image from each of the R, G, B CRT magnified through the lens, reflected through a mirror and projected on a large screen, the consumer's desire for the large screen To meet.

However, this process inevitably causes screen distortion due to optical or mechanical factors. In addition, since the signals generated from the three light sources R, G, and B are projected onto the screen through different paths, the shape of distortion on the screen is different. As a result, the final displayed image has a geometric distortion and color convergence (convergence) does not match.

Conventional analog projection TVs use a method of adjusting convergence by applying a correction current to the convergence yokes of R, G, and B, respectively, to adjust the path of the electron beam. However, this method has a problem that not only a complicated computing device is required to deal with the nonlinear distortion caused by the enlargement of the screen, but also a readjustment after a certain time.

Recently, with the spread of digital TV, research is being conducted to correct optical distortion and color convergence using digital signal processing (DSP) technology. That is, as shown in FIG. 2, inverse transformation of distortion through image signal processing is performed on each of R, G, and B so that the finally corrected image is displayed. That is, the R digital image processor 201 corrects the image distortion of the R signal, analogizes it in the D / A converter 202, and outputs the analog signal to the projection unit 207. The G digital image processor 203 corrects the image distortion of the G signal, analogizes it in the D / A converter 204, and outputs the analog signal to the projection unit 207. The B digital image processor 205 corrects the image distortion of the B signal, analogizes it in the D / A converter 206, and outputs the analog signal to the projection unit 207.

Here, the image signal processing in the R, G, and B digital image processors 201, 203, and 205 means spatial coordinate transformation, which is called warping. When the coordinate system of the source image is (u, v) and the coordinate system of the target image is (x, y), image warping is performed by a high order polynomial (high order) as shown in Equation 1 below. polynomial).

,

는 폴리노미얼의 계수를 나타내고, N은 폴리노미얼의 차수(order)를 나타낸다. 이때, 상기 폴리노미얼의 차수가 높을수록 보정 가능한 왜곡의 유형이 다양해지는 반면 알고리즘의 복잡도는 기하급수적으로 늘어나게 된다. here,

,

Denotes the coefficient of the polynomial, and N denotes the order of the polynomial. At this time, the higher the order of the polynomial, the more various types of distortions that can be corrected, while the complexity of the algorithm increases exponentially.

Optical distortions that we commonly see include shifting in FIG. 3A, scaling in FIG. 3B, horizontal skew in FIG. 3C, vertical skew in FIG. 3D, keystone in FIG. 3E, and FIG. A tilt of 3f, a pincushion of FIG. 3g, a barrel of FIG. 3h, and the like. The distortion that can be corrected according to the polynomial order is as follows.

.1 primary: shifting, scaling, skew, tilt

Secondary: Keystone

.3 car: pincushion, barrel

However, in order to correct such distortion, polynomials having at least three orders of magnitude or more must be used, and warping must be independently processed for each of R, G, and B, thereby increasing the complexity of the algorithm and the implementation cost.                         

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to divide the phenomenon represented by optical distortion into a problem of color convergence and a geometric distortion, and to process each independently, by using polynomials of least order. Another aspect of the present invention is to provide a method and an apparatus for correcting screen distortion, which reduces an implementation cost.

Screen distortion correction method according to the present invention for achieving the above object,

(a) performing coordinate calculation to correct geometrical distortions commonly included in R, G, and B;

(b) performing R data mapping after calculating coordinates for the convergence distortion correction of the R component using the coordinates calculated in the step (a) as an input variable;

(c) calculating coordinates for correcting convergence distortion of the G component using the coordinates calculated in the step (a) as an input variable, and then performing G data mapping;

(d) performing coordinates for B data after calculating coordinates for correcting convergence distortion of the B component using the coordinates calculated in the step (a) as an input variable; And

And projecting R, G, and B values of the target image output in steps (b) to (d) on the screen.

The step (a) is a step of determining a plurality of control points in the source image and determining the coordinates of the pixel at each control point and the corresponding coordinates in the target image, and pseudo-inverse the coordinate pair of the control point determined in the step Computing the coefficient of the polynomial by calculating a matrix, and applying a polynomial coefficient calculated in the step to a mapping function to perform a coordinate calculation to correct the geometric distortion.

In the reverse mapping in steps (b) to (d), the pixel of the calculated coordinate of the source image is mapped to the coordinate of the corresponding target image through interpolation.

According to an aspect of the present invention, there is provided a screen distortion correcting apparatus, comprising: a first warping coordinate calculation unit configured to receive coordinate values of a target image and perform coordinate calculations to correct geometric distortions commonly included in R, G, and B; An R warping unit which calculates coordinates for convergence distortion correction of the R component by using the coordinates calculated by the 1 warping coordinate calculation unit as an input variable and performs the R data mapping, and the coordinates calculated by the first warping coordinate calculation unit. Using G as a input variable, calculate a coordinate for the convergence distortion correction of the G component, perform a G data mapping, and use the coordinates calculated by the first warping coordinate calculation unit as the input variable. After calculating coordinates for convergence distortion correction, the B warping unit performs B data mapping and the R, G, B values of the target image output from the R, G, and B warping units are projected onto the screen. Is characterized in that it is composed of a projection.

The R warping unit calculates the coordinates for the convergence distortion correction of the R component by applying the coordinates calculated by the first warping coordinate calculator to the secondary polynomial, and the calculated R source image. And an R mapping unit configured to map pixels of the coordinates to coordinates of the corresponding R target image through interpolation.

The G warping unit applies a coordinate calculated by the first warping coordinate calculation unit to a second polynomial to calculate a coordinate for the convergence distortion correction of the G component, and the calculated G source image of the And a G mapping unit configured to map pixels of the coordinates to coordinates of the corresponding G target image.

The B warping unit applies a coordinate calculated by the first warping coordinate calculation unit to a second polynomial to calculate a coordinate for the convergence distortion correction of the B component, and the calculated B source image of the B source image. And a B mapping unit for mapping the pixel of the coordinates to the coordinates of the corresponding B target image through interpolation.

Other objects, features and advantages of the present invention will become apparent from the following detailed description of embodiments taken in conjunction with the accompanying drawings.

Hereinafter, with reference to the accompanying drawings illustrating the configuration and operation of the embodiment of the present invention, the configuration and operation of the present invention shown in the drawings and described by it will be described as at least one embodiment, By the technical spirit of the present invention described above and its core configuration and operation is not limited.

4 is a flowchart illustrating an embodiment of an image warping method according to the present invention. First, in step 401, coordinates of several pixels in the source image and coordinates in the target image corresponding thereto are determined.

That is, as shown in FIG. 5, several control points in the source image are determined, and the desired warping shape is determined by changing the positions of these control points.

) ->

i=1,...,M라 하면 폴리노미얼의 계수

는 다음의 수학식 2와 같이 매트릭스(matrix)의 슈도인버스(pseudoinverse)를 통하여 계산된다.In step 402, the coordinate pair of the control point determined in step 401 is substituted into the mapping function of Equation 1 to calculate the coefficient of the polynomial. To get a unique coefficient here, you need to use more control points than the number of polynomial coefficients. If you use M control points and the polynomial order N = 2,

)->

i = 1, ..., M is the coefficient of the polynomial

Is calculated through the pseudoinverse of the matrix as shown in Equation 2 below.

와 상기 수학식 1을 사용하여 좌표 계산을 수행한다. 여기서 매핑 방향에 따라 순방향 매핑(forward mapping) 방법과 역방향 매핑(backward mapping) 방법으로 구분된다. 상기 순방향 매핑은 소스 영상의 각 화소의 좌표로부터 타깃 영상의 좌표를 계산하고, 역방향 매핑은 타깃 영상의 각 화소 좌표로부터 소스 영상의 좌표를 계산한다. 상기된 수학식 1은 역방향 매핑을 나타내는 식이다. 여기서, 소스 영상의 좌표 (u,v)와 타깃 영상의 좌표 (x,y)를 바꾸면 순방향 매핑 식을 얻는다. In step 403, the polynomial coefficient calculated in step 402

The coordinate calculation is performed using Equation 1 and Here, according to the mapping direction, it is classified into a forward mapping method and a backward mapping method. The forward mapping calculates the coordinates of the target image from the coordinates of each pixel of the source image, and the reverse mapping calculates the coordinates of the source image from each pixel coordinate of the target image. Equation 1 is an expression representing reverse mapping. Here, the forward mapping equation is obtained by changing the coordinates (u, v) of the source image and the coordinates (x, y) of the target image.

Finally, in step 404, image data mapping is performed using the coordinates calculated in step 403. In the case of the forward mapping, a process of mapping each pixel to coordinates of the target image calculated by inputting the coordinates of each pixel of the source image. In general, since the calculated coordinates do not exactly correspond to the pixels of the target image, a method of assigning them to the pixel of the closest position or distributing the pixel values at an appropriate ratio to several pixels around the calculated coordinates is used. In this case, in the case of the forward mapping, an unassigned pixel or a pixel that is overlapped may be generated, and thus it should be supplemented through a separate post-processing process. On the other hand, in the case of reverse mapping, since the coordinates of the source image are obtained from the coordinates of the target image to obtain data, the pixels that are not assigned or the pixels that are overlapped do not occur, but the calculated coordinates of the source image do not coincide with the pixel position. Therefore, the interpolated pixel should be used using an interpolation technique.

In general, an image displayed on a projection TV is displayed by projecting an image signal from each CRT of R, G, and B through a lens, and reflecting through a mirror and projecting onto a screen. In this case, distortion of pincushion, barrel, etc. is included in the magnification process through the lens, and keystone, tilt, etc. are included according to the reflection path and the projection angle in the process of reflection and screen projection through the mirror. In addition, distortion such as skew and shifting can be regarded as distortion caused by the CRT itself. Herein, when the lenses of R, G, and B are assumed to have the same curvature, the distortion caused by the lens may be applied to R, G, and B in the same manner. Thus, distortions such as keystones and barrels do not affect color convergence.

Therefore, the present invention is characterized by dividing image distortion into geometric distortion and color convergence distortion by using this fact and performing spatial transformation with minimum order polynomial for each.

FIG. 6 is a block diagram illustrating a screen distortion correcting apparatus according to the present invention when assuming reverse mapping, and includes first warping coordinates for performing coordinate calculations to correct geometric distortions commonly included in R, G, and B. FIG. An R warping unit 602 that calculates coordinates for correcting the convergence distortion of the R component by using the coordinates calculated by the calculator 601 and the first warping coordinate calculator 601 as an input variable, and then performs R data mapping. ), A G warping unit 603 for performing G data mapping after calculating coordinates for correcting the convergence distortion of the G component using the coordinates calculated by the first warping coordinate calculation unit 601 as an input variable. 1 B warping unit 604 for calculating the coordinates for convergence distortion correction of the B component using the coordinates calculated by the warping coordinate calculation unit 601 and performing the B data mapping, and the R, G, In the B warping section (602,03,604) And a projection unit 605 that projects R, G, and B values of the output target image onto the screen.

In the present invention configured as described above, the first warping coordinate calculation unit 601 performs coordinate calculation for correcting geometric distortion commonly included in R, G, and B. At this time, since at least a third order polynomial is required to correct the distortion shown in FIG. 3, a coordinate calculation represented by Equation 3 below is performed.

는 콘트롤 포인트의 좌표 쌍

을 상기된 수학식 2에 대입하여 슈도인버스 매트릭스 연산을 통하여 계산된다. Here, (x, y) represents the coordinates of the target image, (p, q) represents the coordinates of the source image corrected for geometric distortion. Polynomial coefficient

Is the coordinate pair of the control point

Is calculated through the pseudoinverse matrix operation by substituting Equation 2 as described above.

The calculated coordinates are input to the R, G, and B warping units 602 to 604 and used as input variables in calculating the R, G, and B coordinates for color convergence correction.

The R, G, and B warping units 602 to 604 perform data mapping through a warping coordinate calculation unit and an interpolation technique that calculate coordinates by using a secondary polynomial to correct convergence distortion of R, G, and B components, respectively. It consists of a mapping unit. That is, the R warping unit 602 includes an R warping coordinate calculation unit 602a and an R mapping unit 602b, and the G warping coordinate calculation unit 603 includes the G warping coordinate calculation unit 603a and the G mapping unit ( 603b), the B warping coordinate calculation unit 604 is composed of a B warping coordinate calculation unit 604a and a B mapping unit 604b.

That is, the R warping coordinate calculation unit 602a calculates coordinates for correcting convergence distortion of the R component by performing coordinate transformation represented by a second polynomial as shown in Equation 4 below. In this case, the coordinate calculated by the first warping coordinate calculator 601 is used as an input variable in calculating the R coordinate.

The G warping coordinate calculation unit 603a calculates coordinates for correcting convergence distortion of the G component by performing coordinate transformation represented by a second polynomial as shown in Equation 5 below. In this case, the coordinate calculated by the first warping coordinate calculator 601 is used as an input variable in calculating the G coordinate.

The B warping coordinate calculation unit 604a calculates coordinates for correcting convergence distortion of the B component by performing coordinate transformation represented by a second polynomial as shown in Equation 6 below. In this case, the coordinate calculated by the first warping coordinate calculator 601 is used as an input variable in calculating the B coordinate.

,

) , (

,

) , (

,

)는 각각 R, G, B의 소스 영상 좌표를 나타내고, 폴리노미얼 계수

,

,

는 색상 컨버전스 보정을 위한 파라미터로 컨트롤 포인트 좌표 쌍

,

,

으로부터 계산된다. 최종적으로 계산된 소스 영상의 좌 표 (

,

) , (

,

) , (

,

)는 정수 값이 아닌 실수 값이 된다. In the above Equations 4 to 6, (

,

), (

,

), (

,

) Represents the source image coordinates of R, G, and B, respectively, and polynomial coefficients.

,

,

Is a control point coordinate pair as a parameter for color convergence correction

,

,

Is calculated from Coordinates of the final source image (

,

), (

,

), (

,

) Is a real value, not an integer value.

,

) , (

,

) , (

,

)의 화소값을 주위 화소를 이용한 보간 기법을 이용하여 보간하여 해당 타깃 영상의 좌표의 화소값으로 매핑하고 그 결과를 투영부(605)로 출력한다. 예를 들어, R 매핑부(602b)는 계산된 R 소스 영상의 좌표

,

의 화소를 보간을 통해 해당 R 타깃 영상의 좌표에 매핑한다.Accordingly, each of the R, G, and B mapping units 602b to 604b of the R, G and B warping units 602 to 604 may have coordinates of the source image.

,

), (

,

), (

,

) Is interpolated using an interpolation technique using surrounding pixels, and is mapped to pixel values of coordinates of the target image, and the result is output to the projection unit 605. For example, the R mapping unit 602b may calculate the coordinates of the R source image.

,

The pixel of is mapped to the coordinates of the corresponding R target image through interpolation.

(Hardware Minimization Plan)

At this time, assuming that the type of the image distortion is limited to that represented in FIG. 3, Equations 3 to 6 may be shortened as in Equations 7 to 10 below.

Equations 7 to 10 are in the form of removing the third and second terms, respectively. By eliminating polynomial terms, the number of coefficients is reduced, which greatly reduces the complexity of the algorithm and the hardware implementation cost. In addition, the higher order terms have a problem in that the hardware implementation is a big burden by increasing the number of bits of the operation significantly.

As described above, according to the method and apparatus for correcting image distortion according to the present invention, the phenomenon represented by optical distortion is divided into a problem of color convergence and a geometric distortion, and the polynomials of the minimum order required for each are processed separately. By using this, it is possible to correct geometric distortion and color convergence distortion caused by mechanical or optical deformation in display devices such as projection TVs, projectors, monitors, etc., while greatly reducing the implementation cost. In other words, the distortion can be corrected at the minimum cost, so that it can be more competitive in terms of price and performance when applied to projection TVs as well as projectors and monitors.

Those skilled in the art will appreciate that various changes and modifications can be made without departing from the spirit of the present invention.

Therefore, the technical scope of the present invention should not be limited to the contents described in the embodiments, but should be defined by the claims.

Claims (15)

(a) Determine a plurality of control points in the source image, determine the coordinates of the pixel at each control point and the corresponding coordinates in the target image, and then calculate the coefficients of the polynomial from the coordinate pairs of the determined control points. Calculating coordinates of a source image in which geometric distortions, which are commonly included in R, G, and B, are applied to the set reverse mapping function; (b) R pixels at the calculated coordinates of the source image after calculating the coordinates of the source image for the convergence distortion correction of the R component using the coordinates of the source image having the geometric distortion corrected in step (a) as an input variable. Mapping a value to coordinates of a corresponding target image and outputting the mapped value; (c) G pixel at the coordinates of the source image calculated after calculating the coordinates of the source image for the convergence distortion correction of the G component using the coordinates of the source image having the geometric distortion corrected in step (a) as an input variable. Mapping a value to coordinates of a corresponding target image and outputting the mapped value; (d) B pixels at the coordinates of the source image calculated after calculating the coordinates of the source image for correcting the convergence distortion of the B component using the coordinates of the source image having the geometric distortion corrected in step (a) as an input variable. Mapping a value to coordinates of a corresponding target image and outputting the mapped value; And and (e) projecting the R, G, and B values of the target image output in the steps (b) to (d) on the screen. The method of claim 1, wherein step (a) Determining a plurality of control points in the source image and determining coordinates of pixels at each control point and corresponding coordinates in the target image; Calculating coefficients of the polynomial by calculating a coordinate pair of the control point determined in the above step using a pseudoinverse matrix, And applying a polynomial coefficient calculated in the step to a reverse mapping function to calculate coordinates of a source image in which geometric distortion by a lens is corrected. The method of claim 2, wherein the polynomial coefficient calculating step If you use M control points and the polynomial order N = 2,

The coefficient of the polynomial is obtained through the pseudoinverse matrix represented by the following equation.

,

Screen distortion correction method, characterized in that for calculating.

Where (u, v) is the coordinate of the source image and (x, y) is the coordinate of the target image The method of claim 2, wherein the coordinate calculation step Calculated Polynomial Coefficient

To calculate the coordinates of the source image by applying to the following reverse mapping function.

here,

Denotes the coefficient of the polynomial and N denotes the order of the polynomial. The method of claim 1, wherein (b) to (d) And mapping the pixel of the coordinate of the source image calculated for the convergence distortion correction to the coordinate of the corresponding target image through interpolation. The coefficient of the polynomial is calculated from a pair of coordinates of a pixel at a plurality of control points of the source image and a corresponding coordinate in the target image, and applied to a preset reverse mapping function to be common to R, G, and B. A first warping coordinate calculator configured to calculate coordinates of the source image in which the geometric distortion is included; The R pixel value at the calculated coordinates of the source image after calculating the coordinates of the source image for correcting the convergence distortion of the R component by receiving the coordinates of the source image whose geometric distortion is corrected by the first warping coordinate calculator as an input variable R warping unit for outputting the mapping to the coordinate of the target image; The first warping coordinate calculator receives the coordinates of the source image whose geometric distortion is corrected as an input variable, calculates the coordinates of the source image for the convergence distortion correction of the G component, and then calculates the G pixel values at the calculated coordinates of the source image. A G-warping unit configured to map the target image to coordinates of the output; The first warping coordinate calculator receives the coordinates of the source image whose geometric distortion is corrected as an input variable, calculates the coordinates of the source image for the convergence distortion correction of the B component, and then calculates the B pixel values at the calculated coordinates of the source image. B warping unit for mapping the output to the coordinates of the target image; And a projection unit for projecting the R, G, and B values of the target image output from the R, G, and B warping units on a screen. delete delete delete delete The method of claim 6, wherein the first warping coordinate calculation unit If you use M control points and the polynomial order N = 2,

The coefficient of the polynomial is obtained through the pseudoinverse matrix represented by the following equation.

,

Screen distortion correction device, characterized in that for calculating.

Where (u, v) is the coordinate of the source image and (x, y) is the coordinate of the target image. The method of claim 11, wherein the first warping coordinate calculation unit Screen distortion correction, characterized by calculating coordinates (p, q) of a source image in which geometric distortions commonly included in R, G, and B are corrected using a three-dimensional polynomial represented by the following equation. Device.

Here, (x, y) represents the coordinates of the target image, (p, q) represents the coordinates of the source image corrected geometric distortion,

Is the coefficient of the polynomial. The method of claim 12, wherein the R warping portion Coordinates for correcting the convergence distortion of the R component by applying the coordinates (p, q) calculated by the first warping coordinate calculator to the second polynomial represented by the following equation.

R warping coordinate calculation unit for calculating the,

Where

R source image coordinates, polynomial coefficients

Is a control point coordinate pair as a parameter for color convergence correction

Calculated from.) Coordinates of the calculated R source image

And an R mapping unit configured to map pixels of the pixel to coordinates of the corresponding R target image through interpolation. The method of claim 12, wherein the G warping portion Coordinates for correcting convergence distortion of the G component by applying coordinates (p, q) calculated by the first warping coordinate calculator to a second polynomial represented by the following equation.

G warping coordinate calculation unit for calculating the,

Where

G source image coordinates, polynomial coefficients

Is a control point coordinate pair as a parameter for color convergence correction

Calculated from.) Coordinates of the calculated G source image

And a G mapping unit configured to map pixels of the pixel to coordinates of the corresponding G target image through interpolation. The method of claim 12, wherein the B warping portion Coordinates for correcting convergence distortion of the B component by applying coordinates (p, q) calculated by the first warping coordinate calculator to a second polynomial represented by the following equation.

B warping coordinate calculation unit for calculating the,

Where

B source image coordinates, polynomial coefficients

Is a control point coordinate pair as a parameter for color convergence correction

Calculated from.) Coordinates of the calculated B source image

And a B mapping unit configured to map pixels of the pixel to coordinates of the corresponding B target image through interpolation.

Images