KR100301468B1 - Sharpness enhancement circuit of a color image - Google Patents

Abstract

PURPOSE: A circuit for improving color image definition is provided to improve the visual definition for color signal components considering the dull change of boundary portions of the color signal components. CONSTITUTION: A circuit for improving color image definition includes an A/D converter(21) for converting three analog primary color signals output from an RGB matrix to three digital primary color signals, an RGB/YIQ converter(22) for converting the three primary color signals to luminance signals and chrominance signals, a luminance element improving part(23) for filtering the luminance signals output from the RGB/YIQ converter with a high boost filter to output luminance signals of which a high frequency components are accentuated, a chromaticity element improving part(24) for extracting contour information for the chrominance signal output from the RGB/YIQ converter for re-disposing the right and left pixel values horizontally with respect to a reference pixel to obtain sharp transition states of boundary surfaces of the color signals, a YIQ/RGB converter(25) for converting the luminance signal output from the luminance element improving part and the chrominance signal output from the chromaticity element improving part to the three primary color signals, and an D/A converter(26) for converting the three digital primary color signals output from the YIQ/RGB converter to three analog primary color signals to output to an image signal output part(17).

Description

SHARPNESS ENHANCEMENT CIRCUIT OF A COLOR IMAGE}

The present invention relates to a technique for improving the sharpness of a digital color image in order to improve the image quality of a TV. In particular, the problem of luminance signal processing by adding transient response components is solved. Taking into account that the boundary of the color signal component has a dull change, the vertical contour of the chroma signal is extracted to replace the dull change of the chroma signal component with a sharp change to improve the visual clarity of the color signal component. The present invention relates to an image enhancement circuit.

1 is a block diagram of a color image processing circuit according to the prior art, and as shown therein, an image signal detector 11 for detecting a composite image signal CV from an input signal; A low pass filter 12A for low luminance filtering the composite video signal CV at a predetermined cutoff frequency (0 to 3.1 MHz) to detect a luminance signal; A low pass filter 12B for low frequency filtering the composite video signal CV at a predetermined cutoff frequency to detect a carrier color signal; A delay unit 13 for delaying and outputting the luminance signal output from the low pass filter 12A by a time for which the chroma signal component is delayed in the color signal demodulator 15 to be described later; An image quality improving unit 14 which receives the luminance signal output from the delay unit 13 and improves image quality by a processing method by picking; A color signal demodulator (15) for receiving a carrier color signal output from the low pass filter (12B) and demodulating the color difference signals (R-Y) and (B-Y); The three primary color signals R, red, green, and blue are received by receiving the luminance signal Y output from the image quality improving unit 14 and the color difference signals RY and BY output from the color signal demodulator 15. G), (B) RGB matrix 16 to convert; The red, green, and blue three-primary signals (R), (G), (B) is composed of a video signal output unit 17 for amplifying and outputting at a level suitable for display on the CRT, its operation The explanation is as follows.

The low pass filter 12A performs low pass filtering at a predetermined cutoff frequency (0 to 3.1 MHz) on the composite color signal outputted from the image signal detector 11, that is, the composite video signal CV, and detects a signal having a luminance component. The low pass filter 12B performs low pass filtering at another predetermined cutoff frequency (3.1-4.1 MHz) to detect the carrier color signal component.

The luminance signal output from the low pass filter 12A is delayed for a predetermined time (550 ns to 850 ns) through the delay unit 13 and then supplied to the image quality improving unit 14. To compensate for the time difference between the color signal demodulator 15 and the time required for the process of demodulating the carrier color signal into the color difference signal RY and BY components, as compared with the time required for the luminance signal processing. to be.

The image quality improvement unit 14 improves the image quality by a picking technique using the image quality control voltage CTL_V. That is, an overshoot is added to the contour portion of the luminance signal to make the image more clear.

The luminance signal Y output from the image quality improvement unit 14 and the color difference signals RY and BY output from the color signal demodulation unit 15 are red, green, and blue primary colors signals of the RGB matrix 16. R), (G), and (B), which are amplified and output to a level suitable for driving the CRT by the image signal output unit 17 again.

However, in the color image processing circuit according to the prior art, the transient response component is added to the contour portion in order to improve the image quality in the image quality improvement unit. In this case, there is a problem that is very sensitive to noise, and furthermore, the transient response component It was exaggerated and there was a defect that the curvature of the color was outstanding in the outline part.

Accordingly, a technical problem of the present invention is to convert R, G, and B primary colors signals output from the image signal processing circuit into digital signals, and then the luminance signals are sharpened for the image signals using an unsharp masking filter. In the present invention, a sharpness enhancement circuit of a color image which extracts a vertical outline of a chromaticity signal component, rearranges a pixel determined as a pixel on the outline, converts it into an analog signal, and outputs the converted analog signal.

1 is a block diagram of a color image processing circuit according to the prior art.

Figure 2 is a block diagram of an embodiment of a sharpness enhancement circuit of a color image according to the present invention.

FIG. 3 is a detailed block diagram illustrating an embodiment of the chromaticity element improving unit of FIG. 2. FIG.

4 is an exemplary diagram in which an ideal step edge is convolved with a Gaussian function.

(a) is a graph of an image function showing a smooth step edge.

(b) is a graph of the first modified image function.

(c) is a graph of the second modified image function.

*** Description of the symbols for the main parts of the drawings ***

11: video signal detector 12A, 12B: low pass filter

13 delay unit 14 image quality improvement unit

15: color signal demodulator 16: RGB matrix

17: video signal output unit 21: A / D converter

22: RGB / YIQ converter 23: luminance element enhancement unit

24: chromaticity enhancement unit 25: YIQ / RGB converter

26: D / A converter 31A, 31B: low pass filter

32A, 32B: Cutoff Calculator 33A, 33B: Threshold Adjuster

34A, 34B: contour detector 35: adder

36: binary image signal cleaning unit 37: horizontal transient improvement unit

37A: Edge Pixel Detector 37B: Pixel Relocator

37C: Step Edge Inserter

FIG. 2 is a block diagram illustrating an example of a circuit for improving a sharpness of a color image to achieve an object of the present invention. As shown in FIG. 2, an analog three-primary color signal output from an RGB matrix 16 is converted into a digital three-color signal R. A / D converter 21 for converting into (G) and (B); An RGB / YIQ converter 22 for converting three primary color signals R, G, and B into luminance signals Y, color difference signals I, and Q; A luminance element enhancement unit 23 for filtering the luminance signal Y output from the RGB / YIQ converter 22 with an unsharp masking filter and outputting a luminance signal Y 'with a high frequency emphasis; Extract the contour information of the color difference signals (I) and (Q) output from the RGB / YIQ converter 22, so that the pixel values of the left and right horizontally in the horizontal direction centering on the reference pixel so that the boundary of the color signal becomes a sharp transition state. Chromaticity element enhancement unit 24 for rearranging the; The three primary color signals R, red, green, and blue of the luminance signal Y 'output from the luminance element enhancement unit 23 and the color difference signals I' and Q 'output from the chromaticity element enhancement unit 24 are red. A YIQ / RGB converter 25 for converting '), (G'), (B '); The digital three primary color signals R ', (G'), and (B ') output from the YIQ / RGB converter 25 are converted into analog three primary color signals (R), (G) and (B), and then an image signal. The D / A converter 26 output to the output unit 17 side will be described in detail with reference to FIGS. 3 and 4 attached to the operation of the present invention.

The analog three primary color signals output from the RGB matrix 16 are converted into digital three primary color signals R, G, and B by the A / D converter 21 and then to the RGB / YIQ converter 22 again. By the brightness signal (Y), the color difference signal (I, Q) is converted, where the determinant for converting the RGB signal to the YIQ signal is as follows.

In the Y, I, and Q signals converted by Equation 1, Y, which is a luminance signal component, is filtered by a high boost filter (HBP) in the luminance element enhancement unit 23, and the high frequency is emphasized therefrom. A luminance signal Y 'is outputted, and the design formula of this high-pass emphasis filter HBP is as follows.

HBP = (A) original-Lowpass

Equation 2 is commonly referred to as an unsharp masking filter. In the present invention, the two-dimensional spatial unsharp filter (hef) for improving the sharpness of the luminance signal is represented by the following equation.

Where α is a parameter that determines the resulting shape of the masking filter in the range 0 = alpha <= 1. The high-pass emphasized image G _h (j, k) can be obtained by convolution of the unsharp filter hef and the original image F (j, k) for emphasizing the high range as in the following equation.

The degree of high frequency emphasis can be changed by changing the parameter α value of the unsharp filter hef.

In enhancing the sharpness of the chromaticity component output from the RGB / YIQ converter 22, that is, the color difference signals I and Q, the chromaticity element enhancer 24 adds an over-shoot component as in the prior art. Rather than using the peaking method of, the method of changing the edge of the color signal component more precisely is applied.

That is, by extracting the respective vertical contour information for the color difference signals I and Q and rearranging the horizontal pixel values of the pixel determined as the contour, the dull transition state of the color has a sharp horizontal transient characteristic and ultimately visual The sharpness is improved, and the processing thereof will be described in detail with reference to FIG. 3.

It is necessary to rearrange the pixels of the color contours in order to improve the distortion of the shape generated in the process of combining the unnecessary jitter and the color signal of the color edge portion. To this end, first, the input color difference signals I and Q pass through the low pass filters 31A and 31B so as to avoid sensitivity due to noise, and the frequency cut-off characteristics of the low pass filters 31A and 31B are It is expressed as a convolution equation as shown in [Equation 5], and the mask equation of the two-dimensional space filter is expressed as in [Equation 6].

The threshold values T of the I and Q signals respectively set by the threshold adjusters 33A and 33B are parameters for changing the sensitivity of the contour detectors 34A and 34B. As a result, the pixel is rearranged at a high value position causing sufficient color transition, and the sharpness is improved, and the pixel is not rearranged at a position where the color change is not so severe, thereby maintaining the current state.

In order to enable the threshold value T to be appropriately set, the cutoff calculators 32A and 32B calculate the cutoff values for the video signals of I and Q using the following equation.

Here, H is a mask for finding the correlation in the vertical direction, and G is an absolute value of the two-dimensional convolution of the matrixes F and H of the input image. k _x becomes " 1 " when only the vertical edge information is obtained. The boundary value T is the sum of the row vectors plus all the elements of each row of g (rr, cc) divided by the number of elements of the g (rr, cc) matrix as shown in [Equation 10] above. Obtain by multiplying by 4 ". Here, one pixel located at the outermost part of the image is not considered in calculating the boundary value because the 3 × 3 vertical mask is convolved in the m × n size image matrix.

The edge of the given image is found by the obtained boundary value, and the process of finding the contour information is as follows.

The color difference signals I, Q and the boundary values T passing through the low pass filters 31A and 31B are input to the I signal vertical contour detector 34A and the Q signal vertical contour detector 34B, respectively, and are thus applied to the first step. The planarization process for the image is performed. The flattening of the image is performed by first order derivative of the input image with Gaussian function and by convolution.

The two-dimensional Gaussian function is expressed by the following Equation 11, and the first and second derivatives thereof are obtained as shown in Equations 12 and 13.

4 (a)-(c) show the shape when the ideal step edge is convolved with a Gaussian function. Here, the presence and location of the edges are represented by peak and zero-crossing, respectively. In fact, the first derivative of the Gaussian convolved image function is expressed by Equation (14).

It can be seen that the first derivative of the Gaussian function and the convolutional image function are equivalent as shown in Equation 15 below.

Therefore, the image signal can be combined in one convolution in one dimension through the flattening and detection steps by the Gaussian filter. In other words, by convolving with the first derivative of Gaussian, the peak value is found and compared with the previously calculated boundary value T. As a result, when it is larger than the boundary value T, the pixel on the edge is determined.

The image signals obtained by the contour detectors 34A and 34B are binary image signals composed of "0" and "1". Pixels exceeding the threshold value T are white (logical value: 1), and other portions are shown with a black background (logical value: 0).

Subsequently, a new binary image signal is generated by adding the edge image signals of the color difference signals I and Q through the adder 35, and the binary image signal cleaning unit 36 performs the binary image signal cleaning to make the information more useful. A morphology process is performed, which is a process for finding accurate edge pixels by removing independent pixels that are separated from the acquired edge image signal.

By way of example, the removal method of an independent pixel is simply realized by substituting the value of the pixel with " 0 " when all eight neighboring pixels are " 0 " as shown in Equation 16 below.

Thereafter, the horizontal transient improvement unit 37 sets the reference point of the pixel to be repositioned by comparing the original video signal with the binary image signal from which unnecessary independent pixels are removed through the above process. When a reference pixel is determined to replace the boundary of the color signal with a sharp transition state, the pixel values of the left and right are rearranged in the horizontal direction around the pixel.

That is, if any pixel A (i, j) is determined as an edge pixel by the edge pixel detector 37A, the pixel relocator 37B causes the pixel A (i, j) to be expressed as in Equation 17 below. ), The neighboring pixel values in the horizontal direction are rearranged in an appropriate range.

In Equation 17, b is the number of pixels to be transitioned.

The width control signal CTL_W supplied to the steep edge inserter 37C is a signal for adjusting the range of pixels to be rearranged based on the edge pixels of the I and Q signals obtained by the edge pixel detector 37A. .

In this way, the color difference signal (I ', Q') in which the horizontal transient phenomenon of the chromaticity signal component is improved and the luminance signal (Y ') in which the high range is emphasized are converted into red, green, and blue primary colors signals (YIQ / RGB converter 25). R '), (G'), (B ') is inversely transformed, and the inverse transform thereof is obtained by the inverse matrix equation of Equation 1 above.

The inverted red, green, and blue three-primary signals (R '), (G'), (B ') are analog three-primary signals (R), (G), (B) by the D / A converter 26 again. ) And then amplified by the video signal output unit 17 to a level suitable for driving the CRT.

As described in detail above, the present invention converts the three primary color signals into digital signals, and then filters the luminance signals with an unsharp masking filter, extracts vertical contour lines for chroma signal components, and rearranges pixels determined as pixels on the contour lines. By converting and outputting to the next analog signal, jitter and distortion at the edge of the color signal are eliminated, thereby improving visual clarity. In addition, when applied to a digital television receiver, such a technology can be simply implemented as a block inside a digital image processing circuit, and when applied to a general analog television receiver, it can be easily implemented by adding it to the rear end of an image processing circuit. It works.

Claims (3)

An RGB / YIQ converter for converting input three primary color signals into luminance signals and color difference signals; Each boundary value is obtained for the color difference signal output from the RGB / YIQ converter, and the edge information is obtained by comparing the peak value obtained by the derivative with the boundary value after flattening the image through the derivative and the convolution process by the Gaussian function. Set the reference pixel of the pixel to be rearranged by comparing the contour detection unit to be found, the binary image signal newly generated by the contour detection unit, and the original image signal, and then the neighboring pixel values in the horizontal direction around the pixel. A chromaticity element enhancer configured to include a horizontal transient improvement portion which is rearranged so that the boundary portion of the color signal is in a sharp transition state; And a YIQ / RGB converter for converting the luminance signal output from the RGB / YIQ converter side and the color difference signal output from the chromaticity element enhancer into three primary color signals of red, green, and blue. Enhancement circuit. 2. The apparatus of claim 1, wherein the chromaticity element enhancer comprises: a respective low pass filter for removing noise components respectively included in the input chrominance signals I and Q; Respective cutoff calculators and boundary value adjusters for obtaining respective boundary values T for the color difference signals I and Q; The low-pass filtered color difference signal (I), (Q) and the boundary value (T) are provided, and the peak value and the boundary value (T) determined by the derivative after flattening the image through the derivative and convolution process by the Gaussian function. An I signal vertical contour detector and a Q signal vertical contour detector for comparing edges to find edge information; An adder for generating a new binary image by adding respective edge images output from the vertical contour detectors; A binary image signal cleaning unit which performs a morphology processing process to remove unnecessary independent pixels among the binary image signals output from the adder; After comparing the binary image signal outputted from the binary image signal cleaning unit and the original image signal, the reference pixel of the pixel to be rearranged is set, and the neighboring pixel values in the horizontal direction with respect to the pixel are rearranged to an appropriate range. A sharpness enhancement circuit for a color image, characterized by comprising a horizontal transient improvement unit which makes a boundary of a signal become a sharp transition state. An A / D converter for converting an analog three primary color signal output from the RGB matrix into a digital three primary color signal; Each boundary value is obtained for the chrominance signal output from the A / D converter side, and the edge information is compared by the peak value obtained by the derivative and the edge value by flattening the image through the derivative and the convolution process by the Gaussian function. A reference pixel of a pixel to be rearranged by comparing an original image signal with a binary image signal newly generated by the contour detection unit, and then setting adjacent pixel values horizontally around the pixel. A chromaticity element enhancer configured to include a horizontal transient improvement portion which is rearranged in a range so that the boundary portion of the color signal is in a sharp transition state; And a D / A converter for converting the luminance signal output from the A / D converter side and the digital three primary color signal output from the chromaticity element enhancement unit into an analog three primary color signal and outputting the analog signal to the image signal output unit. Sharpness improvement circuit of color image to say.