KR100249652B1 - Apparatus for eliminating pulse noise - Google Patents

Abstract

Disclosed is a pulsed noise removing apparatus capable of removing pulsed noise corresponding to the magnitude of the pulsed noise included in an image signal. The pixel detector inputs an image signal in units of frames to detect current pixels to detect pulsed noise, and horizontal and vertical adjacent pixels, which are adjacent to the current pixel, both horizontally and vertically. The pulse noise detector examines the correlation between the horizontal and vertical adjacent pixels and the current pixel, and detects the pulsed noise in response to the checked correlation. The compensation pixel calculator calculates a compensated pixel value of the current pixel from vertically adjacent pixels, and the pixel selector selects the compensated pixel value or the current pixel as the current pixel from which the pulsed noise has been removed in response to the detection result of the pulse noise detector. Will print By checking the correlation between the current pixel and the adjacent pixel corresponding to the width of the pulse noise, the width of recognition of the pulsed noise is widened, and the neighboring pixels are used to compensate for the pixel value with the noise, thereby preventing the noise from being displayed. have.

Description

Pulsed noise canceller

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for removing pulsed noise, and in particular, to remove pulsed noise present in an image signal in pixels (or pixels) corresponding to its width, and to recover lost image information. Relates to a removal device.

In an image processing system such as a television or video tape recorder, one of the important problems is the reproduction of clean images free of noise components.

In general, screen images provided from a broadcasting station or a videotape making company are likely to contain unnecessary noise due to the image transmission channel or repeated copying. In addition, noise may be introduced into the image information signal even during operation of the image signal processing system.

Such noise may be included in an image source or may also occur when low quality film is used in an image processing system. Moreover, the image is also affected by weak electromagnetic waves while receiving the television image signal.

Noise is generally classified into impulse noise and line noise, and the impulse noise is further classified into white noise and black noise. These noise components may exist simultaneously in the video signal. Therefore, various measures must be taken into account to effectively reduce these noise components.

FIG. 1 is a block diagram of a conventional pulse noise removing apparatus, and includes a pulse noise detector 110, a low pass filter 120, a delay unit 130, and a switching unit 140.

The pulse noise detector 110 detects the presence or absence of pulse noise from the input image signal, and controls the switching operation of the switching unit 140 according to the detection result of the pulse noise.

The low pass filter 120 filters the input image signal to remove noise included in the image signal, and then outputs the image signal from which the noise is removed to the first terminal 141 of the switching unit 140.

The delay unit 130 delays the input image signal by a predetermined time and then outputs the delayed image signal to the second terminal 142 of the switching unit 140.

The switching unit 140 is input to the first terminal 141 and the second terminal 142 from the low pass filter 120 or the delay unit 130 under the control of the pulse noise detector 110. Output video signal selectively.

When the pulse noise detector 110 detects the pulse noise from the input image signal, the pulse noise detector 110 controls the switching unit 140 to output the interpolated image signal by the low pass filter 120. When the pulse noise is not detected from the image signal, the switching unit 140 is controlled to output an image signal output from the delay unit 130 through the second terminal of the switching unit 140.

By the way, in the above configuration, when pulse noise is detected from the video signal, the interpolated video signal is output simply by using a low pass filter. There is a problem that does not correspond to the size of.

An object of the present invention is to provide a pulsed noise removing apparatus that can remove the pulsed noise corresponding to the magnitude of the pulsed noise included in the video signal.

1 is a block diagram of a conventional pulse noise removing apparatus.

2 is a block diagram of a pulsed noise removing device according to an embodiment of the present invention.

FIG. 3 is a diagram illustrating pixels of a frame for explaining the apparatus illustrated in FIG. 2.

4 is a circuit diagram of a compensation pixel calculator shown in FIG. 2.

FIG. 5 is a block diagram illustrating an example of the pixel detection unit illustrated in FIG. 2.

FIG. 6 is a block diagram illustrating an example of a first, second, three, or four pixel delay unit illustrated in FIG. 5.

FIG. 7 is a block diagram illustrating an example of the pulse noise detector illustrated in FIG. 2.

* Explanation of symbols for main parts of the drawings

110, 204: pulse noise detector 120: low pass filter

130: delay unit 140: switch

200: pixel detector 202: compensation pixel calculator

206: pixel selection unit 300: adder

408, 410: first and second horizontal delay parts

400, 402, 404, 406: first, 2, 3 and 4 pixel delays

420: delay unit 440: control unit

500, 502: first and second comparison units

In order to achieve the above object of the present invention, the present invention, by inputting an image signal in units of frames to detect the current pixel to detect the pulse noise, the horizontal and vertical adjacent pixels respectively adjacent to the current pixel and horizontally Pixel detection means; Pulse noise detection means for examining a correlation between the horizontal and vertical adjacent pixels and the current pixel and detecting the pulsed noise in response to the checked correlation; Compensation pixel calculating means for calculating a compensated pixel value of the current pixel from the vertically adjacent pixels; And pixel selecting means for selectively outputting the compensated pixel value or the current pixel as the current pixel from which the pulsed noise has been removed in response to a detection result of the pulse noise detecting means. .

By checking the correlation between the current pixel and the adjacent pixel corresponding to the width of the pulse noise, the width of recognition of the pulsed noise is widened, and the neighboring pixels are used to compensate for the pixel value with the noise, thereby preventing the noise from being displayed. have.

EMBODIMENT OF THE INVENTION Hereinafter, the structure and operation | movement of a pulsed noise removal apparatus by this invention are demonstrated with reference to attached drawing.

2 is a block diagram of an apparatus for removing pulsed noise according to an exemplary embodiment of the present invention, which includes a pixel detector 200, a compensation pixel calculator 202, a pulsed noise detector 204, and a pixel selector 206. It consists of.

FIG. 3 is a diagram illustrating pixels of a frame for explaining the apparatus illustrated in FIG. 2.

2 and 3, the pixel detection unit 200 inputs an image signal in units of frames through an input terminal IN, and detects pulsed noise of the pixel A among pixels of the input image signal frame. Detects as the current pixel, and detects adjacent pixels D and E spaced apart from the current pixel in the horizontal direction as the first and second horizontal adjacent pixels, and the pixels adjacent to the current pixel A in the vertical direction ( F and G) are detected as first and second vertically adjacent pixels.

, 을 출력하는 증폭기(302)로 구성된다.4 is a circuit diagram illustrating an example of the compensation pixel calculator 202 shown in FIG. 2. As shown, the compensation pixel calculator 202 inputs and adds the first and second vertically adjacent pixels F and G, and outputs an adder 300 and an adder 300 that output the added result. Is amplified by 1/2 times so that the average value of the first and second vertically adjacent pixels,

And an amplifier 302 for outputting

The compensation pixel calculator 202 calculates the compensated pixel value of the current pixel. As shown in FIG. 4, the first and second vertical adjacent pixels F and The average value of G) is output to the pixel selector 206 as a compensated pixel value of the current pixel so that it can be used when the current pixel has pulsed noise.

On the other hand, the pulse noise detector 204 examines the correlation between the horizontal and vertical adjacent pixels input from the pixel detector 200 and the current pixel, and detects the pulse noise of the current pixel in response to the checked correlation. Detect. That is, the pixels A, D, E, F, and G detected by the pixel detector 200 are input, and the input first and second vertical adjacent pixels F and G, the first and the second horizontal, are input. Compare whether the value obtained by multiplying the largest pixel value among the adjacent pixels D and E by M (here, M> 1) or less is equal to or less than the current pixel A, or by using the first and second vertical adjacent pixels F And G) comparing whether the value obtained by multiplying the smallest pixel value among the first and second horizontally adjacent pixels D and E by N (where N <1) is equal to or greater than the current pixel A, and comparing According to the result, the selection signal S is output to the pixel selection unit 206.

That is, if the value obtained by multiplying M by the largest pixel value among the first and second vertically adjacent pixels and the first and second horizontally adjacent pixels is less than or equal to the current pixel A, A selection signal S of 1 'is output; otherwise, a selection signal of' 0 'is output. Alternatively, when the value obtained by multiplying N by the smallest pixel value among the first and second vertically adjacent pixels and the first and second horizontally adjacent pixels is equal to or greater than the current pixel A, the selection signal S of '1'. Otherwise, it outputs a selection signal of '0'. Where M and N are values that can be adjusted.

The pulse noise detector 204 checks a correlation between pixels horizontally and vertically adjacent to the current pixel in order to detect pulsed noise of the current pixel. This is because the current pixel on the screen has a high correlation coefficient between adjacent pixels. In addition, the width of the pulse noise is not limited to one pixel, and here three pixels are set as thresholds. At this time, if the correlation between D, E, F, and G and A is significantly low, the current pixel is regarded as pulsed noise.

The pixel selector 206 transfers the compensated pixel value output from the compensation pixel calculator 202 or the current pixel A output from the pixel detector 200 to the first and second terminals 208 and 210, respectively. Input, and switching operation is controlled in response to the selection signal S to compensate one of the signals input to the first and second terminals 208 and 210 by selectively removing the pulsed noise through the output terminal OUT. Output as the current pixel. That is, the pixel selector 206 selects and outputs a compensated pixel value input to the second terminal 210 in response to the selection signal S of '1' and outputs the selection signal S of '0'. When is input, the current pixel input to the first terminal 208 is selected and output as the current pixel from which pulsed noise is removed through the output terminal OUT.

FIG. 5 is a block diagram illustrating an example of the pixel detector 200 illustrated in FIG. 2. As shown, the pixel detector 200 includes first, second, third and fourth pixel delay units 400, 402, 404 and 406, and first and second horizontal delay units 408 and 410.

The first pixel delay unit 400 of the pixel detection unit 200 shown in FIG. 5 delays the image signal input in units of frames through the input terminal IN for a predetermined pixel time, so as to be the first vertical adjacent pixel F. As shown in FIG. The first horizontal delay unit 408 delays the image signal input in units of frames through the input terminal IN for one horizontal synchronizing time and outputs the first horizontal adjacent pixel D. The second pixel delay unit 402 delays the first horizontal adjacent pixel D output from the first horizontal delay unit 408 for the predetermined pixel time and outputs the current pixel A as a third pixel delay. The unit 404 delays the current pixel A output from the second pixel delay unit 402 for the predetermined pixel time and outputs it as the second horizontal adjacent pixel E. The second horizontal delay unit 410 delays the first horizontal adjacent pixel D for one horizontal synchronizing time and outputs the same to the fourth pixel delay unit 406, and the fourth pixel delay unit 406 outputs the second pixel. The image signal delayed by the horizontal delay unit 410 is delayed for the predetermined pixel time and output as the second vertical adjacent pixel G.

FIG. 6 is a block diagram illustrating an example of the first, second, three, or four pixel delay units 400, 402, 404, or 406 shown in FIG. 5. As shown, the first, second, third, or fourth pixel delay units 400, 402, 404, or 406 are the first to X th pixel delay units 422, 424,..., 426, and 428. The pixel delay unit 420 and the control unit 440 are configured.

The first 1 pixel delay unit 422 shown in FIG. 6 inputs an image signal through an input terminal IN, delays 1 pixel, and outputs a 1 pixel delayed image signal to the second 1 pixel delay unit 424. The second one pixel delay unit 424 delays the image signal delayed by one pixel from the first first pixel delay unit 422 by one pixel and outputs the first one to the third one pixel delay unit. That is, the X th pixel delay unit delays the image signal delayed by one pixel from the X-1 th pixel delay unit for one pixel time and outputs the delayed image signal to the X + 1 one pixel delay unit. Meanwhile, the controller 440 may input all the outputs of the first to X th pixel delay units 422, 424,..., 426, and 428 and selectively output the terminal in response to the delay control signal K. FIG. Output through OUT.

As a result, the first, second, third or fourth pixel delay units 400, 402, 404 or 406 shown in FIG. 5 are able to automatically determine the pixel delay period according to the delay control signal K. As shown in FIG. In the above description of the present invention, K = 2 is set, but it can be appreciated that the value may be other than that.

FIG. 7 is a diagram illustrating an example of the pulse noise detector 204 shown in FIG. 2. As shown, the pulse noise detector 204 includes first and second comparators 500 and 502, an OR gate 520 and first, 2, 3, 4, 5, 6, 7 and 8 multipliers. 504, 506, 508, 510, 512, 514, 516 and 518.

The first, second, third, and fourth multipliers 504, 506, 508, and 510 shown in FIG. 7 input D, E, F, and G from the pixel detector 200 shown in FIG. 5, respectively, to multiply M. The result of multiplication is then output to the first comparator 500. The first comparator 500 compares the current pixel with the largest value among the multiplied results. If the largest value is less than or equal to the current pixel, the first comparator 500 generates '1', otherwise generates '0' to generate the OR gate 520. ) Similarly, the fifth, sixth, seventh and eighth multipliers 512, 514, 516, and 518 input D, E, F, and G from the pixel detector 200 shown in FIG. The result is output to the second comparator 502. The second comparator 502 compares the smallest value with the current pixel among the multiplied results, generates '1' if the smallest value is greater than or equal to the current pixel, and generates an OR gate 520 otherwise. )

The OR gate 520 performs an OR on the comparison results output from the first and second comparison units 500 and 502, respectively, and outputs the OR result as the selection signal S to the pixel selection unit 206.

As described above, according to the present invention, the correlation between the current pixel and the adjacent pixel is increased by checking the correlation between the current pixel and the adjacent pixel, and the pixel value having the noise is determined using the adjacent pixels. Compensation can prevent the noise from being displayed.

Although this invention was demonstrated concretely by the said Example, this invention is not restrict | limited by this, A deformation | transformation and improvement are possible within the normal knowledge of a person skilled in the art.

Claims (7)

Pixel detection means (200) for inputting an image signal in units of frames to detect a current pixel to detect pulsed noise and horizontal and vertical adjacent pixels respectively adjacent to the current pixel horizontally and vertically; Pulse noise detection means (204) for checking the correlation between the horizontal and vertical adjacent pixels and the current pixel and detecting the pulsed noise in response to the checked correlation; Compensation pixel calculation means (202) for calculating a compensated pixel value of the current pixel from the vertically adjacent pixels; And pixel selecting means 206 for selectively outputting the compensated pixel value or the current pixel as the current pixel from which the pulsed noise has been removed in response to a detection result of the pulse noise detecting means 204. Pulsed noise canceller characterized in that. 2. The apparatus of claim 1, wherein the pulse noise detecting means 204 multiplies M (where M &gt; 1) with the largest pixel value of the horizontal and vertical adjacent pixels, and determines whether the multiplied value is equal to or less than the current pixel. And a first comparing means (500) for comparing and outputting the compared result as the sensing result. The pulse noise detecting means 204 multiplies the smallest pixel value among the horizontal and vertical adjacent pixels by N (where N &lt; 1), and the multiplied value is equal to or greater than the current pixel. And second comparing means (502) for comparing the results and outputting the compared result as the sensing result. 2. The pulse noise detecting means (204) of claim 1, wherein the pulse noise detecting means (204) multiplies the largest pixel value among the horizontal and vertical adjacent pixels by M (here, M> 1), and the multiplied value is equal to or less than the current pixel. First comparing means 500 for comparing the; Second comparing means (502) for multiplying the smallest pixel value among the horizontal and vertical adjacent pixels by N (where N &lt; 1) and comparing the multiplied value by more than the current pixel; And a logical sum means (520) for ORing the output of the first comparing means (500) and the output of the second comparing means (502) and outputting the detected result as the detection result. The pixel detection means of claim 1, further comprising: first pixel delay means (400) for delaying the image signal input in units of frames for a predetermined pixel time and outputting the first vertical adjacent pixel (F); First horizontal delay means (408) for delaying the image signal input in units of frames for one horizontal synchronizing time and outputting the first horizontal adjacent pixel (D); Second pixel delay means (402) for delaying said first horizontally adjacent pixel (D) for said predetermined pixel time and outputting it as a current pixel (A); Third pixel delay means (404) for delaying said current pixel (A) for said predetermined pixel time and outputting it as said first horizontally adjacent pixel (E); Second horizontal delay means (410) for delaying the first horizontal adjacent pixel (D) for one horizontal synchronizing time; And fourth pixel delay means 406 for delaying the image signal delayed by the second horizontal delay means 410 for the predetermined pixel time and outputting it as a second vertically adjacent pixel G. The second horizontally adjacent pixels D and E are the horizontally adjacent pixels, respectively, and the first and the second vertically adjacent pixels F and G are the vertically adjacent pixels. Device. 6. The method of claim 5, wherein the first, second, third or fourth pixel delay means 400, 402, 404 or 406 comprise first to X first pixel delay means 422, 424, ..., 426 and 428); One of the outputs of the first to X-th first pixel delay means 422, 424,..., 426, and 428 that is input is selectively responsive to the delay control signal K. (F), a control means 440 for outputting as a second horizontal adjacent pixel D, a current pixel A or a second vertical adjacent pixel G, wherein the first first pixel delay means 422 The input image signal is delayed for one pixel time, and the X-th pixel delay means delays the image signal input from the X-1 first pixel delay means for one pixel time and outputs it to the X + 1 one-pixel delay means. Pulsed noise removing device characterized in that. The apparatus of claim 1, wherein the compensation pixel calculator (202) outputs an average value of the vertically adjacent pixels as a compensated pixel value of the current pixel.

Images