KR100252949B1 - scan converter - Google Patents

Abstract

PURPOSE: A scan converter circuit is provided to improve an image quality in using an inter-interpolation and an intra-interpolation method by using an intra-interpolation method in case of a large movement and an inter-interpolation method in case of no movement. CONSTITUTION: A memory unit(20) comprises field memories(13,12,11) to save a present field and a previous field and line memories to delay data outputted from field memories by one line. A subtracting unit(30) calculates a difference between each pixel data. A filter unit(40) removes a noise from a value calculated by the subtracting unit(30). A comparing unit(50) compares a value filtered by the filter unit(40) and calculated by the subtracting unit(30) with a designated value. An interpolation unit(60) outputs a final interpolation data value with selecting an interpolation method in accordance with an output from the comparing unit(50).

Description

Scan converter circuit

The present invention relates to a scan converter circuit for converting interlaced field data into progressive frame data in a TV set or the like.

In recent years, standard TV (SDTV) and the like that have been spotlighted to display only progressive scan frame data. Therefore, if the source is interlaced field data, it should be converted into progressive scan frame data.

At this time, the most commonly used technique for interlaced-sequential scanning conversion is the inter-interpolation method of interpolating the immediately previous field data between the current field line data as shown in FIG. 1 and the current field as shown in FIG. 2. Line interpolation, that is, an intra-interpolation method that interpolates between current field line data with a median value between two lines vertically.

While these methods have the advantage of being able to construct very simple hardware, the inter-interpolation method of FIG. 1 has a larger gap between the top field and the bottom field as the motion increases, thus reducing The image quality is severely deteriorated, such as the screen being distorted in a large part after interpolation, and since the intra-interpolation method of FIG. 2 is forcibly generated, the image quality deteriorates significantly after interpolation of the still part.

Accordingly, in order to solve this problem, US Patent Nos. 5,027,201 and 5,159,451 disclose a technique for detecting motion and converting interlaced scanning into sequential scanning according to the presence or absence of motion. However, these methods have a problem in that the performance of image quality is excellent, but it uses a large amount of field memory and is composed of complex hardware to increase the manufacturing cost of the entire circuit.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to detect the motion of an image and use an intra-interpolation method when the motion is large, and use an inter-interpolation method when there is no motion. A scan converter circuit converts data into progressive scan frame data.

It is another object of the present invention to detect motion of an image and to determine the direction of motion when the amount of motion is small, and to interpolate and interpolate field data by sequentially combining inter-interpolation and intra-interpolation. To provide a scan converter circuit for converting to.

1 illustrates an example of inter-interpolation in a conventional scan converter.

2 illustrates an example of intra-interpolation in a conventional scan converter.

3 is a configuration block diagram of a scan converter circuit according to the present invention;

4 is a diagram illustrating an example of a process for motion detection in a subtractor of FIG. 3;

FIG. 5 is a diagram illustrating an example of a process for detecting the direction of motion in the subtractor of FIG. 3. FIG.

Explanation of symbols for main parts of the drawings

11, 12, 13: field memory 20: memory

21 to 28: line memory 30: subtraction part

40: filter part 50: comparison part

60: interpolator

In accordance with another aspect of the present invention, a scan converter circuit includes a field memory for storing a current field, a previous field, and a previous field, a subtractor for calculating a difference between pixel data, and a value of the subtractor. A comparator which detects the directionality of the motion when the presence or absence of the motion and the amount of the motion is small compared to the predetermined Munter values, and an interpolation part which creates and outputs new pixel data values by an interpolation method determined according to the comparison result Characterized in that configured to include.

The present invention is characterized by obtaining a value for detecting the presence or absence of movement by subtracting pixel data of a previous previous field from current field data.

The present invention obtains a value for detecting the presence or absence of the up direction by subtracting the pixel data value of the current field from the pixel data of the previous field in the up direction, and determining the pixel data value of the current field from the pixel data of the previous field in the down direction. It is characterized by finding the value of detecting the presence or absence of the down direction by subtracting the absolute value.

The present invention is characterized in that it comprises a filter for filtering a value for detecting the presence or absence of motion to remove noise.

The present invention is characterized in that there is no motion if the value for detecting the presence or absence of the motion is less than or equal to a predetermined threshold Min, and the pixel data value of the previous field is used as a new pixel data value.

According to the present invention, if the value for detecting the presence or absence of motion is greater than or equal to a predetermined threshold value Max, the motion is determined to be very large, and the new pixel data value is set as an intermediate value between pixel data of two lines before and after the current field. It is done.

The present invention determines that there is a small amount of movement when the value for detecting the presence or absence of the movement is greater than a predetermined liter value Min and less than the threshold value Max, and when the direction of the movement is determined while detecting the directionality of the movement, m It is characterized by normalizing with '= m-Min and R = Max-Min.

According to the present invention, when the value for detecting the presence of the up direction is smaller than the value for detecting the presence of the down direction, it is determined that the movement is in the up direction, and only when the value for detecting the presence of the up direction is smaller than the predetermined liter value Dir. The final decision is that the movement is in the up direction, multiply the middle value of the pixel data of two lines before and after the current field by the normalized value (m '/ R), and the pixel data value of the previous line after the previous field In addition, a new pixel data value is obtained by multiplying the normalized value (1-m '/ R) and adding the multiplied value of m' / R.

According to the present invention, when the value for detecting the presence or absence of the down direction is smaller than the value for detecting the presence or absence of the direction, the motion is determined to be in the down direction, and the motion for detecting the presence or absence of the down direction is compared with the predetermined liter value Dir. The final decision is made down direction, multiply the middle value of the pixel data of two lines before and after the current field by the normalized value (m '/ R), and add the pixel data of the previous line of the previous field to the pixel data value of the previous field. A new pixel data value is obtained by multiplying a normalized value (1-m '/ R) and adding the multiplied value of m' / R.

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

3 is a block diagram illustrating a scan converter circuit according to the present invention.

Referring to FIG. 3, field memories 13, 12, and 11 storing current fields, previous fields, and previous fields, and a plurality of delays of data output from each of the field memories 13, 12, and 11 by one line. The memory unit 20 composed of the line memory of the memory unit 20, the subtraction unit 30 for calculating the difference between the pixel data output through each line memory of the memory unit 20, the subtraction unit 30 detects the presence or absence of motion Filter unit 40 for removing noise of the calculated value m, the value m filtered by the filter unit 40 and the two values calculated to detect the direction of movement in the subtractor 30 ( a comparator 50 for comparing up and down with a predetermined Munter value, and an interpolator 60 for selecting an interpolation method according to the output of the comparator 50 and outputting a final interpolated data value. do.

In the present invention configured as described above, when data is input, the data of the current field Fn is stored in the field memory 13, the data of the previous field Fn-1 is stored in the field memory 12, and the previous field is stored in the field memory 11. The data of (Fn-2) is stored sequentially.

At this time, three line memories 21 to 23 are sequentially connected to the field memory 11, and in the field memory 11, the data c1 of the current line of the previous field Fn-2 is converted into a line memory ( 21 outputs one line delayed data c2, line memory 22 outputs two line delayed data c3, and line memory 23 outputs three line delayed data c4. Two line memories 24 and 25 are sequentially connected to the field memory 12. In the field memory 12, the data b1 of the current line of the previous field Fn-1 is connected to the line memory 24. Outputs one line delayed data b2 and the line memory 25 outputs two line delayed data b3. Three line memories 26 to 28 are sequentially connected to the field memory 13. In the field memory 13, the data a1 of the current line of the current field Fn is connected, and the line memory 26 is 1. The line delayed data a2 is output, the line memory 27 outputs two line delayed data a3, and the line memory 28 outputs three line delayed data a4.

The subtraction unit 30 has these values a1 to a4, b1 to b3, and c1 to c4 to calculate a value m for detecting the presence of motion and a value up and down for detecting the direction of movement.

In this case, it is assumed that the new pixel data N is created between the a2 line and the a3 line of the current field Fn.

First, the subtractor 30 calculates m as shown in FIG. 4 for motion detection. 4 illustrates an example in which a current field Fn is a top field, a previous field Fn-1 is a bottom field, and a previous field Fn-2 is a top field. In order to obtain the m value, m1, m2, m3, and m4 are first obtained as in Equation 1 below.

이다.

to be.

Then, the largest value m of the four values m1, m2, m3, and m4 is output.

That is, m is obtained as MAX (m1, m2, m3, m4), and the larger the m value, the larger the motion.

In this way, the presence or absence of movement is detected by subtracting the pixel data of the previous previous field Fn-2 from the data of the current field Fn. If the current field is the top field, the difference between the pixel data of the current field and the pixel data of the previous top field is calculated. If the current field is the bottom field, the pixel data of the current field and the pixel data of the previous bottom field are calculated. Motion is detected by calculating the difference. In order to view the surroundings in consideration of noise, for example, four pixels are respectively subtracted and compared with the previous pixel in the vertical direction, and the largest value thereof is output.

In addition, the subtractor 30 calculates up and down values as shown in FIG. 5 shows an example in which the current field Fn is a top field, the previous field Fn-1 is a bottom field, and the previous previous field Fn-2 is a top field.

In order to calculate the up and down values, first, d1 to d6 and u1 to u6 are obtained as in Equation 2 below.

Similarly, the up value is obtained as MAX (u1, u2, u3, u4, u5, u6) and the down value is obtained as MAX (d1, d2, d3, d4, d5, d6).

The up and down thus obtained are used to improve the image quality when the amount of motion is very small, and the existence of the motion is determined by the m value.

By the way, when there is noise in the image, even though there is no motion, the value of m increases and it is considered that there is motion. Therefore, in order to prevent such a case, the m value is passed through the filter unit 40 for median filtering and output to the comparison unit 50.

The comparison unit 50 compares the m value passing through the filter unit 40 and the up and down values output from the subtraction unit 30 with a predetermined threshold values of Min, Max, and Dir. Determine whether it is directional.

1) If m ≤ Min

In this case, since there is no motion, the interpolator 60 generates new pixel data by inter-interpolation using the value of the previous field as it is under the control of the comparator 40. That is, the output of the interpolator 60 becomes b2, and this value becomes N value in FIGS. 4 and 5.

2) when m ≥ Max

를 새로운 픽셀 데이터 값(N)으로 출력한다.In this case, since the motion is large, the intra-interpolation method as shown in FIG. 2 is used. That is, the interpolator 60 is an intermediate value of two lines before and after the vertical direction of the current field Fn under the control of the comparator 50, that is,

Outputs a new pixel data value (N).

3) In case of Min 〈m 〈Max

In this case, since there is a small amount of movement, it is determined whether the movement is directional by comparing the up and down values with the threshold value Dir, and outputs a control signal according to the determination result.

First, the up and down values are compared to determine the direction of movement.

If the up value is less than the down value (up> down), it is determined that the motion of the image is upward, and the variable down_value = 0. Then, the up value is compared with the threshold value Dir to determine if the motion is really upward. If the up value is less than the threshold Dir (up <Dir), then the image is really moving upwards, so set the variable up_value = up, otherwise (up <Dir), if there is no motion, the variable up_value = 0 Release.

On the other hand, if the up value is greater than the down value (up &gt; down), it is determined that the motion of the image is downward, and the variable up_value = 0 is set. The down value is then compared to the threshold value Dir to determine if the motion is really downward. If the value of down is less than the threshold Dir (down <Dir), then the image is really moving downward, so specify the variable down_value = down; otherwise (down> Dir), if there is no movement, the variable down_value = 0 To be specified.

The reason for this is that if the maximum value (up) of u1 to u6 is smaller than the predetermined liter value Dir, the information indicates that the image is moving upward, and the maximum value (down) of d1 to d6 is the liter value Dir. Smaller is because it indicates that the image is moving downward.

Therefore, when the direction of movement is determined, normalize to m '= m-Min and R = Max-Min, and output the values (m', R), up_value, and down_value to the interpolator 60. do.

The interpolation unit 60 obtains and outputs the new pixel data (N) as shown in Equation 3 since the motion direction of the image is upward when up_value> 0.

가 되고, m'/R이 1이면 움직임이 아주 큰 경우이므로,

가 된다. 또한, 움직임의 방향이 위쪽이니까 b3가 새로운 픽셀 데이터(N)쪽으로 움직일 확률이 크므로 새로운 픽셀 데이터(N)는 b3의 영향을 더 받도록한다.Here, m '/ R is a ratio indicating how close the image is to the moving side, and 1-m' / R is a ratio indicating how close to the still side. That is, if m '/ R is 0, there is no motion,

If m '/ R is 1, the motion is very large,

Becomes In addition, since the direction of movement is upward, the probability of b3 moving toward the new pixel data N is large, so that the new pixel data N is further affected by b3.

On the other hand, if down_value> 0, the direction of motion of the image is downward, so that the interpolator 60 obtains and outputs new pixel data N as shown in Equation 4 below.

In this case, since the direction of movement is downward, there is a high probability that b1 moves toward the new pixel data N, so that the new pixel data N is further affected by b1.

가 된다.In addition, if up_value ≤ 0 or down_value ≤ 0, the direction of motion cannot be detected. If the motion is still, m '/ R is 0, so output (N) = b2, and if the motion is very large, m' / R. Is 1

Becomes

As described above, according to the scan converter circuit according to the present invention, if the motion is detected and the motion is large, intra pixel interpolation method is used to generate the new pixel data, and if there is no motion, the new pixel data is converted into a sequential scan frame. By judging the direction of the motion, new pixel data is generated in consideration of the size and direction of the motion, and converted into a sequential scanning frame, so that the image quality generated by the inter-interpolation and intra-interpolation methods improves deterioration and the motion is small. Image quality can be improved and circuit manufacturing costs can be reduced by using fewer field memories and hardware.

Claims (8)

A field memory for storing the current field, the previous field, and the previous previous field, respectively; A line memory for storing line data output from each of the field memories; The absolute value is obtained by subtracting a plurality of pixel data in the vertical direction before and after the vertical direction of the current field data output from each of the line memories and subtracting the plurality of pixel data in the vertical direction. m) is determined and output, and a plurality of pixels before and after the vertical direction of the current field (or the previous field) in a plurality of pixels data in the vertical direction of the previous field (or the previous previous field) in the up direction output from each line memory. The absolute value is obtained by subtracting each data value, and the largest value is output as a value for detecting the presence of the up direction (up), and in the plurality of pixel data before and after the vertical direction of the previous field (or the previous previous field) in the down direction. Absolute values are obtained by subtracting a plurality of pixel data values before and after the vertical direction of the current field (or the previous field), and then downward the largest value among them. A subtractor for outputting a value (down) for detecting the presence or absence, A comparison unit comparing the output of the subtraction unit with predetermined thresholds to detect the presence or absence of movement and the degree of movement when the movement is small; And an interpolation unit for selecting an interpolation method according to the output of the comparison unit, and determining and outputting a new fax data value. The method of claim 1, wherein the subtraction unit And if the current field is a top field, calculating a difference between pixel data of the current field and pixel data of the previous top field to obtain a value (m) for detecting the presence or absence of movement. The method of claim 1, wherein the subtraction unit And if the current field is a bottom field, calculates a difference between pixel data of the current field and pixel data of a previous bottom field to obtain a value (m) for detecting the presence or absence of movement. The method of claim 1, wherein the subtraction unit And a filter unit for filtering a value (m) for detecting the presence or absence of motion to remove noise. The method of claim 1, wherein the comparison unit If the value up for detecting the presence of the up direction is smaller than the value for detecting the presence or absence of the down direction, it is determined that the movement is in the up direction, and a predetermined liter value Dir for determining the value up for detecting the presence of the up direction again. Scan converter circuit, characterized in that it is configured to finally determine that the movement is in the up direction only in comparison with The method of claim 5, wherein the interpolation unit If the comparison unit determines that the direction of motion is the up direction, the middle value of the pixel data of the two lines before and after the current field is multiplied by the normalized value (m '/ R), and the pixel data value of the previous field is the line after the previous field. And multiplying the normalized value (1-m '/ R) by adding the pixel data of and then adding the multiplied value of m' / R to obtain a new pixel data value. The method of claim 1, wherein the comparison unit If the value for detecting the presence or absence of the down direction (down) is smaller than the value for detecting the presence or absence of the direction (up), it is determined that the movement is in the down direction, and the predetermined value Dir for determining the value (down) for detecting the presence of the down direction again and Scan converter circuit, characterized in that it is finally determined that the movement is in the down direction only in comparison. The method of claim 7, wherein the interpolation unit When the comparison unit determines that the direction of movement is the down direction, the middle value of the pixel data of two lines before and after the current field is multiplied by a normalized value (m '/ R), and the pixel data value of the previous field is the previous line of the previous field. And multiplying the normalized value (1-m '/ R) by adding the pixel data of to obtain the new pixel data value by adding the multiplied value of m' / R.

Images