KR100226836B1 - Apparatus for field format conversion - Google Patents

Abstract

An image format conversion apparatus for converting interlaced scanning into a sequential scanning method. In particular, an edge detection unit which detects an edge on a pixel-by-pixel basis by receiving current field image data and 2-field delayed field image data, and determines a detected edge movement And a motion detector for detecting motion by comparing the difference between the brightness components of the two field images with a preset threshold value only when it is determined that there is motion on the edge detected by the edge detector, and the motion information detected by the motion detector. A horizontal and vertical extension unit for extending the motion data in a horizontal and vertical direction, and a deinterlacing unit for performing line compression after interpolating the data according to the degree of movement of the data output from the horizontal and vertical extension units. To detect the edge on a pixel-by-pixel basis It does not make an error to judge that there is motion in the background image without motion by judging whether it is motionless, or to judge that there is no motion, and after detecting the motion of the edge, it calculates and compares the difference between two original image fields again. In the process of comparing the edges, it is possible to prevent the detection of motion or the removal of motion.

Description

Video format converter

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for converting interlacing, which is an NTSC method, into a progressive scan. In particular, an image format conversion for performing deinterlacing by determining a motion using an edge portion of an image is performed. Relates to a device.

A typical television video signal adopts an interlaced scanning method in which two fields form one frame, thereby performing compression of a frequency band transmitted. In recent years, PCs and high-definition televisions attempt to display the interlaced scans on the screen. However, since PCs or high-definition televisions usually display in a sequential manner, in order to display a conventional interlaced scan, an image line that is not present in a conventional interlaced scan must be arbitrarily made to be sequentially scanned, which is de-interlaced. -interlacing). To do this, the motion of the image must be detected first.

That is, in the conventional technique of converting interlaced scanning into sequential scanning, motion detection is detected by using a difference of field images having two frame differences or by using multiple fields.

For example, the technique disclosed in US Pat. No. 5,488,422 finds motion with only the brightness component of a field image having only two frame differences, that is, if the brightness component with two frame differences is above a certain value, there is a motion and a certain value. Since it is determined that there is no motion below, it indicates that there is a considerable part of the motion even in an image having substantially no motion. Thus, a processor, such as a filter, is needed to remove this error in the post-processing part after finding the movement.

In addition, the technique disclosed in U.S. Patent No. 5,027,201 requires a large number of field memories because several field images are required to find motion. That is, since the motion detection is performed by using the pixel difference between the diagonals of the five field images and the pixel difference between the three fields with four field memories, basically, many field memories are used.

As described above, when the motion detection is performed using only two field differences as in the former case, many errors occur. Otherwise, when the motion detection is performed using several field images as in the latter case, many field memories are required. There was a problem.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to extract an edge by extracting each edge component only from two field images using two field memories, thereby accurately detecting the motion of an object. The present invention provides a format conversion apparatus.

A feature of the image format conversion apparatus according to the present invention for achieving the above object is an edge detection unit for detecting the edge in the pixel unit by receiving the current field data and two field delayed field data and determines the movement of the detected edge And a brightness component detector which compares current field data with two field delayed field data, obtains and latches the difference between the brightness components of the two images, and the brightness component detector only when it is determined that there is motion in the edge detected by the edge detector. A motion detector for detecting motion by receiving the difference between the brightness components of the two images from a second threshold and a horizontal and vertical extension for expanding the motion data in the horizontal and vertical directions according to the motion information detected by the motion detector; Section of the data output from the horizontal and vertical expansion units. After the data interpolation according to the degree jikim may consists of parts that perform the de-interlacing a line compression.

1 is a block diagram of an image format conversion apparatus according to the present invention

FIG. 2 is a detailed block diagram of the edge detector of FIG. 1. FIG.

3 is a detailed block diagram of the edge detector of FIG.

4 is a detailed block diagram of the horizontal extension of FIG.

FIG. 5 is a detailed block diagram of vertical expansion in the vertical expansion and data sample part of FIG. 1; FIG.

FIG. 6 is a detailed block diagram of a data sample in the vertical expansion and data sample unit of FIG. 1. FIG.

Explanation of symbols for the main parts of the drawings

101: A / D converter 102, 103: field memory

104: edge detector 105: subtractor

107: buffer 108: switching unit

109: motion detection unit 110: horizontal expansion unit

111: vertical expansion and data sample section 112: 2: 1 line compression section

113: D / A converter 201,202: edge detector

203: Subtractor

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a block diagram illustrating an overall configuration of an image format conversion apparatus according to an embodiment of the present invention, in which a complex video signal (interlaced video signal) input using a comb filter (not shown) is converted into a luminance signal Y and a color. The signal C is separated and the separated luminance signal Y is deinterlaced.

That is, referring to FIG. 1, an analog / digital (A / D) converter 101 for converting the separated luminance signal Y into a digital signal is converted into digital in the A / D converter 101. First field memory 102 for storing 263 lines H in the interlaced scan image output, data connected to the first field memory 102 and delayed by one field in the first field memory 102, that is, 262 lines The second field memory 103 for storing the first field image data output from the second field memory 103 and the third field image data output through the A / D converter 101 in pixel units. After detecting an edge, the edge detector 104 obtains a difference between two images on a pixel basis and outputs the switching control signal, and the first field image data output from the second field memory 103 and the A / D converter ( Third field video day output through 101) The subtractor 105 for outputting the difference between the two images, the buffer 107 temporarily latching the output of the subtractor 105, and the switching control signal output from the edge detector 104 are switched to the buffer ( The switching unit 108 is selectively connected to the 107, and the motion detection unit 109 for detecting the motion of the image by comparing the data output through the switching unit 108 and a predetermined threshold value is configured. Here, reference numeral 106 denotes a line to which the switching unit 18 is connected when the edge detector 104 determines that no edge is detected or no movement is detected even when the edge is detected. The line 106 may be grounded. This line can be used to input meaningless data.

As shown in FIG. 2, the edge detector 104 detects an edge in units of pixels in a third field image, an edge detector 202 in which edges are detected in pixels in a first field image, and the two edges. The subtractor 203 obtains the difference between the edges detected by the detectors 201 and 202 and outputs the switching control signal.

The two edge detectors 201 and 202 have the same configuration, and one of them is shown in detail in FIG. 3. Referring to FIG. 3, a first line memory 301 for delaying one line of field image data, a second line memory 302 for delaying one line delayed by the first line memory 301 again, is inputted. Field image data, that is, buffers 303, 304, and 305 which sequentially latch from the first pixel to the third pixel of the third line, and a multiplier 306 that multiplies the second pixel output from the buffer 304 by an arbitrary constant α. An adder 307 that adds a first pixel and a third pixel output from the buffers 303 and 305, and a second pixel output from the multiplier 306, and a first line of the second line output from the first line memory 301. Buffers 308, 309 and 310 sequentially latching from the first pixel to the third pixel, a multiplier 311 that multiplies the first pixel output from the buffer 310 by an arbitrary constant α, and three outputs from the buffer 308 Th pixel A multiplier 312 multiplying a constant α, buffers 313, 314, 315 sequentially latching from the first pixel to the third pixel of the first line output from the second line memory 302, in the buffer 314 A multiplier 316 multiplying the second pixel output by an arbitrary constant α, an adder that adds a first pixel and a third pixel output from the buffers 313 and 315, and a second pixel output from the multiplier 316 ( 317), the first pixel of the third line output from the buffer 305 and the first pixel of the second line output from the multiplier 311 and the first pixel of the first line output from the buffer 315. An adder 318 that adds the third pixel of the third line output from the buffer 303 and the third pixel of the second line output from the multiplier 312 and the first output of the buffer 313. Of the line An adder 319 for adding the second pixel; It consists of an absolute value part 325 which takes an absolute value in a result. Here, the arbitrary constant α is 2 as an example, but may vary depending on the design purpose. That is, the larger the value of the constant α, the thicker the edge is, so that the edge is roughly detected, and the smaller the value of the constant α is, the more accurately the edge is detected.

According to the present invention configured as described above, the composite video signal (interlaced scanning video signal) is separated into the luminance signal Y and the color signal C, and the separated luminance signal Y is inputted to the A / D converter 101. It is made of a signal of 8-bit digital brightness components.

The first field image data digitized by the A / D converter 101 is stored in the field memory 102 storing 263 lines, and when the second field image data is output through the A / D converter 101, the field memory is output. The second field image data is stored in the field memory 103 which is output from the 102 and stores 262 lines, and the second field image data is stored in the field memory 102. Then, when the third field image data output through the A / D converter 101 is input to the field memory 102 and output to the edge detector 104 and the subtractor 105, the third field image data stored in the field memory 103 is stored. The first field image data is also output to the edge detector 104 and the subtractor 105.

The subtractor 105 obtains the difference between the first field image and the third field image, that is, the difference between the brightness components and latches the buffer 107 until the edge is detected by the edge detector 104.

Meanwhile, the edge detector 201 of the edge detector 104 detects an edge in pixels in the third field image, and the edge detector 202 detects an edge in pixels in the first field image. Here, the configuration of the edge detector 201 and the edge detector 202 is the same, one of which is shown in FIG. In this case, since the first line memory 301 and the second line memory 302 each store one line in the field image, three lines are required to detect an edge in pixel units.

That is, as shown in FIG. 3, the data of the first line is stored in the first line memory 301, and when the data of the second line comes in, it is output to the second line memory 303 and then to the second line memory 302. The second line of data is stored in the first line memory 301. At this time, when the data of the third line comes in, the first pixel, the second pixel, and the third pixel of the first line stored in the second line memory 302 are serially latched in the buffers 313, 314, and 315. That is, the first pixel output from the second line memory 302 is latched in the buffer 313, and when the second pixel is output, the first pixel is output from the buffer 313, latched in the buffer 314, and the second pixel is output. When the pixel is latched to the buffer 313, the third pixel is output, the first pixel is output from the buffer 314, latched to the buffer 315, and the second pixel is output from the buffer 313 to the buffer 314. The third pixel is latched in the buffer 313. Therefore, the first pixel is buffered in the buffer 315, the second pixel is buffered in the buffer 314, and the third pixel is buffered in the buffer 313, respectively.

In the same way, the first pixel of the second line output from the first line memory 301 is latched in the buffer 310, the second pixel is latched in the buffer 309, and the third pixel is latched in the buffer 308. do. In addition, the first pixel of the third line is latched in the buffer 305, the second pixel is latched in the buffer 304, and the third pixel is latched in the buffer 303.

At this time, 1 pixel of the first line is represented by x01, 2 pixels, and 3 pixels are represented by x02 and x03, respectively. Similarly, 1 to 3 pixels of the second line are x11, x12 and x13, and 1 pixel of the third line. To 3 pixels are represented by x21, x22, and x23, and an arbitrary constant α is assumed to be 2. If the second pixel x12 of the second line is detected as an edge in the embodiment, the multiplier 306 outputs the second pixel data x22 of the third line output from the buffer 304. Multiplied by 2 and output to the adder 307, the multiplier 311 multiplies the first pixel data (x11) of the second line output from the buffer 310 by 2 to output to the adder 318, multiplier 312 Multiplies the third pixel data (x13) of the second line output from the buffer 308 by 2 to output to the adder 319, and the multiplier 316 outputs the second pixel of the first line output from the buffer 314. The data x02 is multiplied by 2 and output to the adder 317.

The adder 317 adds the first pixel x01 and the third pixel x03 of the first line output from the buffers 313 and 315 and the second pixel 2x02 output from the multiplier 316 (x01). + 2x02 + x03). The adder 307 adds the first pixel x21 and the third pixel x23 of the third line output from the buffers 303 and 305 and the second pixel 2x22 output from the multiplier 306 (x21). + 2x22 + x23). The adder 318 is the first pixel (x21) of the third line output from the buffer 305 and the first pixel (2x11) of the second line output from the multiplier 311 and the buffer 315 The first pixel x01 of the first line to be output is added (x01 + 2x11 + x21). The adder 319 is the third pixel (x23) of the third line output from the buffer 303 and the third pixel (2x13) of the second line output from the multiplier 312 and the buffer 313 The third pixel (x03) of the first line output from is added (x03 + 2x13 + x23).

The addition results of the adders 307, 317, 318 and 319 are latched in the respective buffers 320, 321, 322 and 323, and then output to the adder 324 at the same time. In this case, when the pixel has a -weight, it is enough to take a two's complement (2's) from the corresponding buffer and output the result. That is, when the buffer 321 takes two's complement to the output of the adder 317, the output of the buffer 321 becomes -x01-2x02-x03. In addition, when the buffer 322 takes two's complement to the output of the adder 318, the output of the buffer 322 becomes -x01-2x11-x21. Then, when the output of the buffers 320, 321, 322, 323 is added in the adder 324, (-x01-2x02-x03) + (x21 + 2x22 + x23) + (-x01-2x11-x21) + (x03 + 2x13 + x23 ) The absolute digitizer 325 detects an edge after taking an absolute value to the addition result of the adder 324.

In this case, when the absolute value is taken by the absolute value unit 325 to the addition result of the adder 324, the following equation (1) is obtained.

In addition, the absolute value unit 325 compares the calculated absolute value with a certain threshold to determine whether the corresponding pixel, that is, the second pixel x12 of the second line, is an edge, and outputs a logic signal accordingly. Here, when the threshold value is set to 255, if the absolute value of the addition result is 255 or more, it is determined as an edge and 1 is output. If it is 255 or less, it is determined that it is not an edge and 0 is output.

Thus, the detected edge movement is determined by the output of the edge detector 201 detecting the edge of the third field image using the subtractor 203 and the output of the edge detector 202 detecting the edge of the first field image. Find the difference signal.

That is, when the second pixel x12 of the second line is taken as an example, if the edge detector 201 outputs 1 and the edge detector 202 also outputs 1, the second pixel x12 of the second line is an edge. However, since there is no change in the third field and the first field, there is no movement. Therefore, if the output of the subtractor 203 is 0, there is no motion at the detected edge, and if 1, there is motion at the detected edge.

If the edge detector 201 outputs 1 and the edge detector 202 outputs 0, or the edge detector 201 outputs 0 and the edge detector 202 outputs 1, the second pixel of the second line is output. Since (x12) is an edge and has a different value in the third field and the first field, it can be seen that the output of the subtractor 203 is 1 and there is a movement.

As described above, the edge detector 104 detects an edge on a pixel-by-pixel basis by using a correlation of 3x3 pixels and determines a movement of the detected edge.

At this time, the signal output from the subtractor 203 is a switching control signal of the switching unit 108. When 0 is output, it is determined that there is no movement on the detected edge, and when 1 is output, it is determined that there is motion on the detected edge. .

Therefore, the switching unit 108 is switched to 106 lines when 0 is output from the edge detector 104. The signal input to line 106 may be zero, or may be any arbitrary value that is not meaningful. On the other hand, when 1 is output from the edge detector 104, the switching unit 108 is connected to the buffer 107. In this case, since the difference between the brightness components of the first field and the third field image is latched in the buffer 107, when the switching unit 108 is connected to the buffer 107, the difference between the brightness components of the two fields is detected. Is input to the unit 109.

The motion detector 109 detects a moving object in the image by giving a threshold value to the image of the brightness component in detecting the motion. That is, if the difference between the brightness components inputted through the switching unit 108 is greater than or equal to a value, it is determined that there is motion and if it is less than or equal to a value, there is no motion, and the result is output. At this time, a value between the maximum value and the minimum value may be divided by the degree of movement. In the present invention, the maximum value and the minimum value are separated in three stages, which can reduce the bandwidth for the motion data value in the next block.

On the other hand, since it is difficult to detect motion with only one pixel, the horizontal expansion unit 110 and the vertical expansion unit and the data sample unit 111 expand the movement data in the horizontal direction and the vertical direction to see the surrounding situation, and deinterlacing the data for deinterlacing. Interpolate

As shown in FIG. 4, the horizontal expansion unit 110 expands the motion data in the horizontal direction, and expands it in units of 8 pixels as the maximum value of the motion data values. That is, data of 1 pixel to 8 pixels are sequentially input to each buffer 400 to 407 and output to the maximum value detector 408. The maximum value detector 408 finds the maximum value among these values and performs vertical expansion and data samples. Output to the unit 111. Here, the maximum value is output because the maximum value has the most movement.

As illustrated in FIG. 5, the vertical expansion and data sample unit 111 sets four pixels as the highest value among four pixels in the vertical direction. That is, since data should be performed in line units, one line of data is sequentially stored in the line memory 502 via the line memories 500 and 501, and two lines of data are stored in the line memory 501 via the line memory 500. ), And three lines of data are stored in the line memory 500. The maximum value detector 503 compares the pixel of the fourth line currently input with each pixel output from the line memories 500 to 502 in the vertical direction and outputs the highest value.

The second field data output from the first field memory 102 and the first field data output from the second field memory 103 may be used to display data not present in the field image according to the maximum value, that is, the degree of movement. Perform interpolation of the data you produce. That is, as shown in FIG. 6, data is generated by combining pixels of the second field output from the first field memory 102 and pixels of the first field output from the second field memory 103, depending on whether the motion is performed or not. Serve

). 감산기(603)는 상기 제산기(602)의 출력에서 첫 번째 필드의 픽셀 a를 빼고(

- a), 곱셈기(605)는 상기 감산기(603)의 결과(

- a)에 움직임 검출상태를 나타내는 계수 α를 곱한다((

- a) * α). 그리고나서, 가산기(606)는 상기 곱셈기(605)의 결과에 첫 번째 필드의 픽셀 a를 더하여 새로운 픽셀(d = (

- a) * α + a) 을 출력한다.For example, a pixel of the first field output from the second field memory 103 is called a, and a pixel of the second field currently output from the first field memory 102 is c, 1H in the delayer 600. If the previous pixel delayed by 1H is called b, the adder 601 adds the pixel c of the currently incoming second field and the delayed pixel c in the 1H delayer 600 (b + c), and the divider 602 is the adder ( The addition result (b + c) of 600 is divided by two (

). Subtractor 603 subtracts pixel a of the first field from the output of divider 602 (

a), the multiplier 605 is the result of the subtractor 603 (

a) is multiplied by a coefficient α representing the motion detection state ((

a) * α). Then, the adder 606 adds the pixel a of the first field to the result of the multiplier 605 to add a new pixel d = (

a) Print * α + a).

)으로 만들어진다. 즉, 움직임이 없으면 α는 0이므로, 가산기(606)의 출력에서 a 픽셀만 남고, 움직임이 있어 α가 1이면

픽셀만 남는다.At this time, the α is the output of the horizontal expansion unit 110, 0 when there is no movement at all, 1 if the output is very large, a new pixel (d) is made of a pixel if there is no movement, the movement is If present, the average value of the two pixels in the second field (

Is made of). That is, if there is no motion, α is 0, so if only a pixel remains at the output of the adder 606, and if there is motion and α is 1,

Only pixels remain.

In this way, the newly generated pixels in the vertical expansion and data sample unit 111 are input to the 2: 1 line compression unit 112 together with the original pixels. Here, since new pixels have been created, they appear to have been converted to sequential scanning data, but since they are twice as wide as the sequential scanning data in time, time compression must be performed. Accordingly, the 2: 1 line compression unit 112 compresses the data at least 1/2 times as a sequential scan-converted data, that is, outputs a luminance signal. To this end, the 2: 1 line compression unit 112 may receive the data output from the vertical expansion and data sample unit 111 as a normal clock and output at least twice as much as the normal clock at the time of output. . The luminance signal converted into the sequential scan by the 2: 1 line compression unit 112 is converted into analog by the D / A converter 113 and output.

As described above, according to the image format conversion apparatus according to the present invention, by detecting an edge on a pixel basis and accurately determining the movement of the detected edge, it is determined that there is motion or there is no movement in the background image without motion. After detecting the edge movement, the difference between the two original image fields is obtained and compared again to prevent the detection of the movement or the removal of the movement portion in the edge comparison process. have. This allows two field memories to be sufficient so that the deinterlacing can be performed accurately while reducing the number of field memories.

Claims (9)

A video format conversion apparatus for digitalizing a interlaced field video signal and performing deinterlacing in a sequential scanning method, A first field memory for delaying and outputting the digitalized field image data by one field; A second field memory configured to delay and output data output by being delayed by one field in the first field memory by one field delay; An edge detector which detects an edge on a pixel-by-pixel basis by receiving field image data currently input and field image data delayed by two fields from the second field memory, and determines movement of the detected edge; A brightness component detector for comparing the currently input field image data with the field image data delayed by two fields in the second field memory to obtain and latch a difference between the brightness components of the two images; A motion detector that detects motion by receiving a difference between brightness components of two images from the brightness component detector only when it is determined that there is motion on the edge detected by the edge detector; Horizontal and vertical expansion units configured to expand motion data in horizontal and vertical directions according to the motion information detected by the motion detection unit; And a deinterlacing unit which performs line compression after interpolating data according to the degree of movement of data output from the horizontal and vertical expansion units. The method of claim 1, wherein the edge detector An edge detector for detecting an edge in units of pixels from currently input field image data; An edge detector for detecting an edge on a pixel-by-pixel basis from field image data delayed by two fields in the second field memory; And an edge motion discriminator for comparing the outputs of the two edge detectors to determine the detected edge motion. The method of claim 2, wherein the edge detector is And an edge is detected by using a correlation of 3x3 pixels. The method of claim 2, wherein the edge detector is An adder that adds weights to the values of the three pixels on the top, bottom, left, and right sides (-x01-x02-x03 + x21 + x22 + x23 -x01-x11-x21 + x03 + x13 + x23); And an absolute value unit which determines an edge by taking an absolute value of the addition result and comparing the threshold value with a preset threshold value. Here, x1, x02, x03 for 1 to 3 pixels on the first line, x11, x12, x13 for 1 to 3 pixels on the second line, and x21 for 1 to 3 pixels on the third line. Defined as x22, x23. The method of claim 1, wherein the horizontal and vertical expansion portion An apparatus for converting video formats, wherein the motion data is extended in a horizontal direction to a maximum value among the motion data values. The method of claim 1, wherein the horizontal and vertical expansion portion And comparing the predetermined pixels in the vertical direction and setting the predetermined pixels to the highest value among them. The method of claim 1, wherein the de-interlacing unit And if the pixel to be interpolated is determined to have no motion, interpolate to the pixel of the previous field. The method of claim 1, wherein the de-interlacing unit If it is determined that the pixel to be interpolated has movement, the average value of the upper and lower pixels of the current field (

And an image format conversion device characterized by interpolation. The method of claim 1, wherein the de-interlacing unit And performing time compression for receiving data as a normal clock and outputting data at least twice as large as the normal clock.

Images