KR100224860B1 - Vertical interpolation method and apparatus and still video formation method and apparatus using the same - Google Patents

Abstract

The present invention provides a method and apparatus for interpolating a line of a given video signal and a line, and a method and apparatus for extracting and forming a still image from a video by applying the same. The present invention is to detect the edge portion, in order to solve the step phenomenon or jaggedness of the edge portion caused by the data difference between the fields due to the time difference between two fields, especially when moving video or motion caused by hand shake in interlaced scanning method. Based on the technique of interpolating the edge part in the tilt direction, the interlaced image is converted into a sequential scan image to form a still image of one frame. The present invention is effective when a still image is obtained from a video from a video camera, a camcorder, a TV, or a VCR, displayed on a screen, or output through a printer. The present invention obtains the edge component of the input data, obtains the interpolated state according to the direction of the interpolation to be interpolated from the edge component, and generates interpolated image data in consideration of the edge of the input data according to the interpolated state. In addition, the present invention synthesizes the generated image data and the input data to obtain a still image. The apparatus includes means for detecting an edge component from input data, means for detecting a state indicating an inclination direction from the edge component, and means for outputting an interpolated image by weighted interpolation of the edge components of the input data according to the state of the inclination direction. do. In addition, the present invention includes a synthesizing means for synthesizing the interpolated image and the input data.

Description

Method and device for vertical interpolation of video signal and method and device for forming still image using same

The present invention relates to a vertical interpolation method and apparatus for a given video signal, and to a still image forming method and apparatus using the same. In particular, an edge generated by a time difference between two fields by employing an interpolation method adaptively interpolated according to an edge component is provided. Vertical interpolation method and device of video signal that solves the problem of step or jaggedness of the part, and by applying the interlaced (interlaced) scanning method to convert the moving image to the progressive (sequential) scanning method A method and apparatus for forming a still image.

For example, when one frame is made up of two fields and a still image is obtained by extracting one frame image from a video signal from a video camera, camcorder, TV, VCR or the like in which the video signal scanning method is an interlaced scanning method. When motion occurs in the video due to motion or moving images due to the time difference between the first field and the second field, the difference between each field is prominent in the eye, and the stairs at the edges of the area where the motion occurs Phenomenon or sawtooth phenomenon occurs.

In order to eliminate the difference between the fields of the interlaced video signal generated by such motion, it is necessary to appropriately correct the interlaced video data. This method is called a de-intelace method, which includes a method of generating data of one frame by interpolating lines and lines of one field data, and generating data of one frame using data of two or more fields. There is a way. The present invention aims to improve a method of forming data of one frame by interpolating lines and lines of one field data in the former case.

1 to 4 illustrate a method of interpolating lines and lines of a given image signal as a typical method of the conventional deinterlace method.

FIG. 1 is a diagram for explaining a method of obtaining a new image by interpolating lines and lines of a given video signal of two fields and one frame by a conventional line doubling method, and FIG. 2 by a conventional vertical linear interpolation method. A diagram for explaining a method of obtaining a new image by interpolating lines and lines of a given video signal of two fields and one frame. The common concept of the two methods is to generate one frame of data by interpolating lines and lines for a given video signal and generating another field based on one field, and the reference image (field or frame image with a given video signal). Each method is divided according to how the data is referred to.

As shown in FIG. 1, the line interpolation method by line dubling generates one field as a reference field and another field by line doubling by the line doubler 15. Using the data corresponding to the position of the data, the frame combining unit 18 adds the data 10 of the reference field and the data 12 of the generated field to form an image of one frame. In this case, when data of an arbitrary frame is X, i is a horizontal delay, and j is a vertical delay, even field data is represented by X [i, 2j], and odd field data is represented by X [i, 2j + 1]. The line interpolation method by the line doubling method is represented by X [i, 2j + 1] = X [i, 2j] when the even field is a reference field and the odd field is a field.

In the line interpolation method of the vertical linear interpolation method shown in FIG. 2, one field is used as a reference field and the other field is generated by a vertical linear interpolation method. The reference field is delayed by the line delay unit 25 for a predetermined time. The vertical interpolator 27 receives the delayed value and the reference field and generates the average value of the vertically adjacent reference field values. Next, the data 22 of the field generated by the frame synthesizing unit 28 is combined with the data 20 of the reference field to obtain a new image of a line and a line interpolated. Also in this case, when the even field is referred to as the reference field and the radix field is generated, it is represented by X [i, 2j + 1] = (X [i, 2j] + X [i, 2j + 2]) / 2. do.

3 and 4, the above-described problems of the conventional methods are severely generated in the stepped edge portion, so that the frame image is not natural.

The technical problem to be achieved in the present invention is a method of interpolating a line and a line of a given video signal, by detecting an edge portion having a slope and interpolating a line and a line of a given image signal adaptively according to the slope portion. It is to provide a vertical interpolation method of the video signal to obtain a.

Another object of the present invention is to provide an apparatus for performing the vertical interpolation method of the video signal.

Another technical problem to be solved by the present invention is a still image forming method for obtaining a still image from a given image signal, by detecting an edge portion having a slope of a given image signal and adaptively interpolating according to the slope portion and interpolating the The present invention provides a still image forming method for synthesizing an image signal and the given image signal to obtain an image that makes the image frame more natural.

Another object of the present invention is to provide an apparatus for performing the still image forming method.

FIG. 1 is a diagram for explaining a method of performing line interpolation of a given video signal and a method of obtaining a one-frame video composed of two fields by a conventional line dubling method.

FIG. 2 is a diagram for explaining a method of performing line interpolation of a video signal and a method of obtaining a one-frame image composed of two fields by a conventional vertical linear interpolation method.

FIG. 3 is a diagram illustrating a one-frame image obtained by FIG. 1.

FIG. 4 is a diagram illustrating a one-frame image obtained by FIG. 2.

5 is a block diagram illustrating a vertical interpolation method and apparatus of an image signal of the present invention and a still image forming method and apparatus using the same.

FIG. 6 is a diagram illustrating a one-frame image obtained by FIG. 5 of the present invention. FIG.

FIG. 7 is a diagram for describing an operation of the tilt direction detection unit of FIG. 5.

FIG. 8 is a diagram for describing an operation of a weighted interpolator that performs interpolation according to weights of edge component data of FIG. 5.

<Explanation of symbols for the main parts of the drawings>

100 edge detection, 110 tilt direction detection

120 ... weighted interpolation unit, 130 ... video synthesis unit

710 upper left, 720 upper right

730 lower left, 740 lower right

810 ... First state, 820 ... Second state

830 ... third state.

The present invention, the process of obtaining the edge component of a given video signal to achieve the above technical problem; Obtaining a state to interpolate according to a tilt direction of surrounding image data to be interpolated; And generating image data by interpolating lines and lines of the given image signal according to the state to be interpolated.

In addition, the present invention, means for detecting the edge component from a given video signal in order to achieve the technical problem; Means for detecting a state indicating a slope direction from the edge component; And means for outputting an interpolated image by interpolating while weighting the edge components of the given video signal according to the state indicating the detected tilt direction.

In addition, the present invention, to achieve the above technical problem, the process of obtaining the edge component of a given video signal; Obtaining a state to interpolate according to a tilt direction of surrounding image data to be interpolated; Generating image data by interpolating lines and lines of the given image signal according to the state to be interpolated; And obtaining a still image by synthesizing the generated image data and the input data.

In addition, the present invention, means for detecting the edge component of a given video signal in order to achieve the technical problem; Means for detecting a state indicating a slope direction from the edge component; Means for outputting an interpolated image by interpolating while weighting an edge component of the input data according to a state indicating the detected tilt direction; And synthesizing means for synthesizing the interpolated image and the given image and outputting the synthesized image.

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

A method of interpolating a line and a line of a given video signal of the present invention, and forming a still image using the same, obtains an edge component of a given video signal, that is, input data, as shown in FIG. 5, and detects the detected edge component. And obtaining an inclination direction of the image data to be interpolated from the surrounding area, adaptively interpolating the image data according to the state indicating the inclination direction, and synthesizing the weighted interpolated image data and the input data. To this end, the still image forming apparatus of the present invention is an edge detection unit 100 for detecting an edge component from the input data, a tilt direction detection unit 110 for detecting the inclination direction from the edge component divided into three states, and the detected The weighted interpolator 120 outputs an interpolated image by interpolating the weighted edge components of the input data according to the state divided according to the tilt direction, and the image synthesizer 130 which synthesizes the weighted interpolated image data and the input data. ).

Next, a vertical interpolation method and apparatus of an image signal and a still image forming method and apparatus using the same will be described. In the process of interpolating a line and a line, a video signal to be newly formed by interpolating the line and the line is X, i is a horizontal delay and j is a vertical delay. The data will be referred to as X _re [i, j] and the interpolated data will be referred to as X _gen [i, j].

The edge detection unit 100 uses a conventional edge detection filter as a block for detecting an edge component with respect to the input data X _re [i, j] of the reference field. One of the easiest forms of such an edge detection filter is a high pass filter. That is, the edge detector 100 obtains the edge component data E _re [i, j] with respect to the input data X _re [i, j] and outputs the edge component data to the inclination direction detector 110.

The inclination direction detection unit 110 is a block for obtaining an inclination around data to be interpolated based on the detected edge component data, which is calculated through a microcomputer or inclination direction detection circuit in hardware, and the calculation algorithm is as follows. same.

In the present invention, a given video signal referred to by data to be interpolated, that is, edge component data of a reference field has a configuration as shown in FIG. As shown in FIG. 7, the larger the number n of edge component data of the reference field to be referred to, the more closely the horizontal slope can be detected.

When the reference field corresponding to the periphery of the position to be interpolated is shown in FIG. 7, among the data of the reference field corresponding to the periphery of the position to be interpolated, the reference field is spatially included in the upper left, upper right, lower left and lower right portions. If the sum of the data of the edge component data of E _re [i, j] is the upper left sum (U _l ), the upper right sum (U _r ), the lower left sum (D _l ), and the lower right sum (D _r ), respectively, It is represented by Formula (1)-(4), respectively.

The slope direction of data to be interpolated is obtained by using values of the upper left sum (U _l ), the upper right sum (U _r ), the lower left sum (D _l ), and the lower right sum (D _r ) obtained through the above equations. Use

First, the slope directions of the data to be interpolated are divided into three states using the upper left sum (U _l ), the upper right sum (U _r ), the lower left sum (D _l ), and the lower right sum (D _r ).

State 1: when U _l > (U _r + T) and D _r > (D _l + T),

State 2: when U _r > (U _l + T) and D _l > (D _r + T), and

State 3: Not in State 1 and State 2. Where T is a threshold value and may be appropriately determined by a designer.

In other words, state 1 is greater than the upper left sum 710 is greater than the upper right sum 720 plus the threshold value T, and the lower right sum 740 adds the lower left sum 730 to the threshold value T. In larger cases, it is shown in the state 810 of FIG. State 2 is greater than the upper right sum 720 plus the threshold T to the upper left sum 710, and the lower left sum 730 is greater than the lower T sum to the lower right sum 740. In this case, the state is indicated by state 820 of FIG. 8. State 3 is represented by the state of 830 of FIG. 8 except for the state 1 and the state 2.

Accordingly, when the obtained states 1, 2 and 3 are " states to be interpolated, " the directions of tilt, that is, directions to interpolate, are obtained from the respective states.

The weighted interpolation unit 120 is an interpolation block according to the state to be interpolated obtained through the tilt direction detection unit 110 and performs an interpolation operation according to each state as follows.

The direction and method to be interpolated according to each state obtained through the tilt direction detection unit 110 are determined as shown in FIG. 8. First, in the case of state 3, the linear interpolation method in the vertical direction is applied as in the conventional interpolation method. In 1 and 2, it is determined as interpolation values of input data values of the reference field to which the weight of the edge component data of the upper left and lower right or the upper right and lower left is respectively applied.

Expressed as a formula to clarify the above described contents are as follows.

In the case of state 1, among the data of a given video signal corresponding to the periphery of the position to be interpolated, the average value is proportional to the edge component of the data included in the upper left spatially and is proportional to the edge component of the data contained in the lower right spatially. The mean value of the mean value is given by the following equation (5),

In the case of state 2, among the data of a given video signal corresponding to the periphery of the position to be interpolated, the average value is proportional to the edge component of the data included in the upper right spatially and the edge component of the data contained in the lower left spatially. The mean value of the mean value is given by the following equation (6),

In state 3, interpolation is performed by performing the interpolation between lines by using the method given as the following equation (7) as the average value of two spatially vertically adjacent data among the data of a given video signal corresponding to the periphery of the position to be interpolated. Generated data.

Through the above-described process, the interpolated data X _gen [i, j] of the generation field for the input data X _re [i, j] of the reference field is obtained, and then synthesized with the reference field input data and outputted. As shown in FIG. 5, the inclined portion between the input data and the interpolated data can form one frame data of a still image which is more gentle. The vertical interpolation method and apparatus of the image signal of the present invention and the still image forming method and apparatus using the same can be used in a color video printer which captures and prints a still image of one frame from a TV or other video signal, and also interlaced. The present invention can also be applied to the electronic zoom function of a digital camcorder using a scanning solid state imaging device (CCD).

As described above, the present invention has been described with respect to a two-field one-frame image. However, the present invention may be similarly applied to a case in which two or more fields form an image. In FIG. 7 and FIG. 8, interpolation is made with five peripheral pixels. Although the method of the present invention has been described, since it is obvious that the method and apparatus for vertical interpolation of the image signal and the still image forming method and apparatus using the same can be applied even with three peripheral pixels, the detailed description thereof will be omitted.

As described above, the method and apparatus for forming a still image frame according to the present invention connects a video signal from an interlaced video camera, a camcorder, a TV, a VCR, or the like, which is widely used in homes, to a personal computer. It can be used when a still image is obtained, and the slanted portion of the edge portion generated by the time difference between the fields is adaptively interpolated along the edge slope direction, thereby making the slanted portion smoother.

Claims (12)

In the vertical interpolation method of a video signal, (a) obtaining an edge component of a given video signal; (b) obtaining a state to be interpolated according to a tilt direction of surrounding image data to be interpolated; And and (c) generating image data by interpolating lines and lines of the given image signal according to the state to be interpolated, and obtaining the image data. The method of claim 1, wherein step (b) (b1) Edge component data E _re [i of input data X _re [i, j] contained spatially in the upper left, upper right, lower left and lower right of the data of a given video signal corresponding to the periphery of the position to be interpolated; , meet the upper left, respectively, as equations below, the sum of the data of the j] (U _l), upper right match (U _r), the left lower meet (D _l), and finding the right lower meet (D _r); And

(b2) The upper left sum is greater than a certain constant greater than the upper right sum by using the values of the upper left sum (U _l ), the upper right sum (U _r ), the lower left sum (D _l ), and the lower right sum (D _r ). State greater than a certain constant greater than the lower left sum (state 1), state that the upper right sum is greater than or equal to the constant constant greater than the upper left sum, and the lower left sum greater than or equal to the constant constant greater than the lower right sum (state 2) and state 1 and not state 2 State (state 3) as follows, If U _l > (U _r + T) and D _r > (D _l + T), state 1, If U _r > (U _l + T) and D _l > (D _r + T), state 2, If not in the state 1 and state 2, the vertical interpolation method of the video signal, characterized in that output as the gradient data of the state 3 (where a constant constant T is a threshold value can be taken according to the designer according to the designer). The method of claim 2, wherein step (c) In the case of state 1, among the data of a given video signal corresponding to the periphery of the position to be interpolated, the average value is proportional to the edge component of the data included in the upper left spatially and is proportional to the edge component of the data contained in the lower right spatially. The mean value of the mean value is given by

In the case of state 2, among the data of a given video signal corresponding to the periphery of the position to be interpolated, the average value is proportional to the edge component of the data included in the upper right spatially and the edge component of the data contained in the lower left spatially. The mean value of the mean value is given by

In the state 3, by using the method given as the average value of two spatially vertically adjacent data among the data of a given video signal corresponding to the periphery of the position to be interpolated,

A vertical interpolation method of an image signal characterized by line interpolation. In the apparatus for interpolating lines and lines of a video signal, Means for detecting an edge component from a given video signal; Means for detecting a state indicating a slope direction from the edge component; And And a weighted interpolation means for outputting an interpolated image by interpolating the weighted edge components of the given video signal according to the state indicating the detected tilt direction. The method of claim 4, wherein the state detection means indicating the inclination direction Edge component data E _re [i, j] of input data X _re [i, j] contained spatially in the upper left, upper right, lower left and lower right among data of a given video signal corresponding to the periphery of the position to be interpolated meet each of the upper left as the sum of the data and formulas below (U _l), to obtain a consistent idol (U _r), the left lower meet (D _l), right lower meet (D _r);

The upper left sum is greater than a certain constant than the upper right sum by using the upper left sum (U _l ), the upper right sum (U _r ), the lower left sum (D _l ), and the lower right sum (D _r ), and the lower right sum is more than the lower left sum A state larger than a certain constant (state 1), a right upper sum greater than a constant constant greater than a left upper sum, a lower left sum greater than a constant constant greater than a lower right sum (state 2), and a state other than state 1 and state 2 (state 3) as follows; If U _l > (U _r + T) and D _r > (D _l + T), state 1, If U _r > (U _l + T) and D _l > (D _r + T), state 2, When not in the state 1 and state 2, the vertical interpolation apparatus of the image signal characterized in that output as the gradient data of the state 3 (where a constant constant T is a threshold value can take an appropriate value according to the designer). The method of claim 5, wherein the weighted interpolation means In the case of state 1, among the data of a given video signal corresponding to the periphery of the position to be interpolated, the average value is proportional to the edge component of the data included in the upper left spatially and is proportional to the edge component of the data contained in the lower right spatially. The mean value of the mean value is given by

In the case of state 2, among the data of a given video signal corresponding to the periphery of the position to be interpolated, the average value is proportional to the edge component of the data included in the upper right spatially and the edge component of the data contained in the lower left spatially. The mean value of the mean value is given by

In the state 3, by using the method given as the average value of two spatially vertically adjacent data among the data of a given video signal corresponding to the periphery of the position to be interpolated,

A vertical interpolation apparatus of a video signal, characterized in that the line interpolation. In a method of forming a still image from a given video signal, (a) obtaining an edge component of a given video signal; (b) obtaining a state to be interpolated according to a tilt direction of surrounding image data to be interpolated; (c) generating image data by interpolating lines and lines of the given image signal according to the state to be interpolated; And and (d) synthesizing the generated image data with the given image signal to obtain a still image. 8. The method of claim 7, wherein step (b) (b1) Edge component data E _re [i of input data X _re [i, j] contained spatially in the upper left, upper right, lower left and lower right of the data of a given video signal corresponding to the periphery of the position to be interpolated; , meet the upper left, respectively, as equations below, the sum of the data of the j] (U _l), upper right match (U _r), the left lower meet (D _l), and finding the right lower meet (D _r); And

(b2) The upper left sum is greater than a certain constant greater than the upper right sum by using the values of the upper left sum (U _l ), the upper right sum (U _r ), the lower left sum (D _l ), and the lower right sum (D _r ). State greater than a certain constant greater than the lower left sum (state 1), state that the upper right sum is greater than or equal to the constant constant greater than the upper left sum, and the lower left sum greater than or equal to the constant constant greater than the lower right sum (state 2) and state 1 and not state 2 State (state 3) as follows, If U _l > (U _r + T) and D _r > (D _l + T), state 1, If U _r > (U _l + T) and D _l > (D _r + T), state 2, When not in the state 1 and state 2, the still image forming method characterized in that the output as the tilt direction data of the state 3 (where a constant constant T can take an appropriate value according to the designer as a threshold value). The method of claim 8, wherein step (c) In the case of state 1, among the data of a given video signal corresponding to the periphery of the position to be interpolated, the average value is proportional to the edge component of the data included in the upper left spatially and is proportional to the edge component of the data contained in the lower right spatially. The mean value of the mean value is given by

In the case of state 2, among the data of a given video signal corresponding to the periphery of the position to be interpolated, the average value is proportional to the edge component of the data included in the upper right spatially and the edge component of the data contained in the lower left spatially. The mean value of the mean value is given by

In the state 3, by using the method given as the average value of two spatially vertically adjacent data among the data of a given video signal corresponding to the periphery of the position to be interpolated,

Still image forming method characterized in that the line interpolation. An apparatus for forming a still image from a given video signal, Means for detecting an edge component of the given video signal; Means for detecting a state indicating a slope direction from the edge component; Means for outputting an interpolated image by interpolating while weighting edge components of data of the given video signal according to the state indicating the detected tilt direction; And And a synthesizing means for synthesizing the interpolated image and the given image and outputting the synthesized image. 11. The method of claim 10, wherein the state detecting means indicating the inclination direction Edge component data E _re [i, j] of input data X _re [i, j] contained spatially in the upper left, upper right, lower left and lower right among data of a given video signal corresponding to the periphery of the position to be interpolated meet each of the upper left as the sum of the data and formulas below _(l U), to obtain a consistent idol _(r U), lower left meet _(l D), consistent with the lower right (D _r);

The upper left sum is greater than a certain constant than the upper right sum by using the upper left sum (U _l ), the upper right sum (U _r ), the lower left sum (D _l ), and the lower right sum (D _r ), and the lower right sum is more than the lower left sum A state larger than a certain constant (state 1), a right upper sum greater than a constant constant greater than a left upper sum, a lower left sum greater than a constant constant greater than a lower right sum (state 2), and a state other than state 1 and state 2 (state 3) as follows; If U _l > (U _r + T) and D _r > (D _l + T), state 1, If U _r > (U _l + T) and D _l > (D _r + T), state 2, If not 1 and 2, the still image forming apparatus, characterized in that output as the tilt direction data of the state 3 (where the constant constant T is a threshold value can take an appropriate value according to the designer). The method of claim 11, wherein the weighted interpolation means In the case of state 1, among the data of a given video signal corresponding to the periphery of the position to be interpolated, the average value is proportional to the edge component of the data included in the upper left spatially and is proportional to the edge component of the data contained in the lower right spatially. The mean value of the mean value is given by

In the case of state 2, among the data of a given video signal corresponding to the periphery of the position to be interpolated, the average value is proportional to the edge component of the data included in the upper right spatially and the edge component of the data contained in the lower left spatially. The mean value of the mean value is given by

In the state 3, by using the method given as an average value of two spatially vertically adjacent data among the data of a given video signal corresponding to the periphery of the position to be interpolated,

Still image forming apparatus characterized in that the line interpolation.

Images