JP2560724B2 - Image processing method - Google Patents

Description

The present invention relates to a signal processing method for transmitting or recording an image.

[Conventional technology]

Today's television signals are basically interlaced scanned signals, and one frame of image is
It consists of two Tsuno field paintings, Odd and Even. As a conventional image processing apparatus for image transmission or image recording, an apparatus for recording 1-field information as it is on a magnetic disk is already known.

[Problems to be solved by the invention]

A field image is basically a frame image that is sampled by thinning out one pixel in the vertical (vertical) direction. If this sampling interval does not satisfy the Nyquist condition, the frame image is interpolated from this field image. Even when trying to reproduce, the image quality is deteriorated due to the air cylinder.

As a method for solving this problem, a method of cutting the high frequency band of the frame signal in advance so that the sampling interval of the field image satisfies the Nyquist condition can be considered.
This solution is effective in the case of a still image composed of two field images, frame image Odd and Even.
However, if there is a time difference between the Odd field and the Even field as in a moving image, there will be image movement between the field images, and a frame image will be formed with field images with different contents. When the low-pass operation is applied to the image signal, the image is blurred at the blur portion of the image.

On the other hand, depending on the image, when a field image is formed by thinning out without performing a low-pass operation on the frame image, air silling occurs, but since the field image consists of image information at a certain time, it becomes an image. May be easier to see.

The present invention has been made in view of the above circumstances, and an object thereof is to provide an image processing method for selecting and executing an appropriate process according to the situation.

[Means for solving problems]

An image processing method according to the present invention is a method for extracting a field signal for recording or transmission from a frame signal,
It is determined whether the frame signal stored in the frame memory is a moving image or a still image. If the frame signal is a still image, the frame signal in the frame memory is subjected to a low pass filter operation in the vertical direction and then the vertical signal is applied. Pixels are thinned out in the direction to form the field signal.
It is characterized in that the field signal is formed by thinning out pixels in the vertical direction from the frame signal of the frame memory without performing the filter calculation.

[Action]

As described above, by performing the low-pass filter calculation depending on whether it is a moving image or a still image, it is possible to form a preferable field signal for each case.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 shows its flow chart, and FIG. 2 shows a block diagram of an implementation circuit of the method.

In FIG. 2, 10 is a frame memory having a storage capacity of one frame, and 12 is a frame image stored in the frame memory 10 in the case where the frame image is a still image.・ Pass filter processing is performed, and then one field signal (for example, Odd field signal) is extracted and output.
It is a digital signal processing circuit that extracts and outputs one field signal without performing low-pass filter processing. Reference numeral 14 is a recording / transmission device for recording or transmitting the field signal from the signal processing circuit 12 to a recording medium or a transmission path.

Referring to FIG. 1, first, an Even field image forming a frame image stored in the frame memory 10 and
It is determined whether the frame image is a moving image or a still image by comparing with the Odd field image (S1). A specific example of this determination method will be described later.

When it is determined to be a moving image in S1, the signal processing circuit 12 makes a low-pass signal in the vertical direction of the frame signal of the frame memory 10.
Apply filter operation. The frame signal subjected to the low pass filter operation is once returned to the frame memory 10. Then, when it is determined to be a still image in S1 and after the low-pass filter operation in S2, the signal processing circuit 12
Is a recording / transmission device for extracting one field signal by thinning out the frame signal of the frame memory 10 in the vertical direction.
Send to 14 (S3). The recording / transmission device 14 performs signal processing for recording or transmission on the received field signal, and then
It is recorded on a recording medium such as a magnetic tape or a magnetic disk, or sent to a transmission line such as a telephone line.

FIG. 3 shows ideal characteristics of the low pass filter operation in S2 (that is, the signal processing circuit 12). The horizontal axis represents the number of horizontal scanning lines per image, and the vertical axis represents the passage rate. The field signal bandwidth to the field sampling frequency
Since it needs to be limited to 1/2, it must eventually be band limited to 1/4 of the frame sampling frequency.
As a digital filter with such characteristics,
Various known methods such as FIR (finite time impulse response) and IIR (indefinite period impulse response) may be used.

An example of a method of discriminating between moving images and still images will be described with reference to FIGS. 4, 5, and 6. In FIG. 4, 20 is an A / D converter for converting an analog input video signal into a digital signal, 22 is a frame memory capable of storing one frame of video signal, and 24 is read from the frame memory 22. A D / A converter that converts the digital signal to an analog signal, 26 is a CPU that controls the A / D converter 20, the frame memory 22 and the D / A converter 24, and performs the operation of moving image / still image discrimination. is there. Specifically, the CPU 26 determines that the absolute value of the density difference between two vertically adjacent pixels in the first field is less than or equal to a first predetermined value, and that the pixels in the second field are the two pixels. The number of pixels matching the condition that the absolute value of the density difference between the pixel located between the two and one of the two pixels is the second predetermined value or more is counted for one screen, and the count value is the predetermined number or more. When it is, it is determined to be a moving image.

Figure 5 shows the coordinates (x, y) in the frame memory 22.
FIG. 3 is a diagram showing the relationship between the density (luminance) value I (x, y) of the pixel on this coordinate. In this case, (i, 2j + 1) is the first
Pixels in the (odd) field, and (i, 2j) are pixels in the second (even) field.

FIG. 6 (1) shows an example of movement of a rectangular object, and shows a case where it is at the position of the solid line R1 during the first field scanning and moves to the position of the broken line R2 during the second field scanning. The scanning line for the rectangle R1 in the first field is as shown in FIG. 6 (2), and the scanning line for the rectangle R2 in the second field is as in (3). FIG. 6 (4) is a combination of (2) and (3). Here, the shaded portion of the solid line and the shaded portion of the dotted line are high density portions, and the other portions are low density portions.

When the above judgment condition of the CPU 26 is mathematically expressed, it can be expressed as follows.

| I (i, 2j) −I (i, 2j + 2) | ≦ δ1 (1) | I (i, 2j) −I (i, 2j + 1) | ≧ δ2 (2) where δ1 and δ2 are set in advance. Δ1
Corresponds to the first predetermined value, and δ2 corresponds to the second predetermined value. Equation (1) has the correlation in the y direction in the first field. That is, only when the density difference between two vertically adjacent pixels in the first field (in the second field) is δ1 or less, the moving image determination is based. Expression (2) is an expression for determining the movement of the object on the screen in the lateral (x) direction.

Actually, when determining the moving image / still image, if the above equations (1) and (2) are both satisfied, the evaluation value C (i, 2j + 1) of the pixel is set to “1”, and otherwise. Is
Set to “0”. Then, when ΣC (i, 2j + 1) is equal to or larger than a certain threshold δ3, it is determined to be a moving image, and when it is less than δ3, it is determined to be a still image. δ3 is set to, for example, 5 to 10% of the number of dots on the entire screen.

Both Expression (1) and Expression (2) are satisfied, for example, when the change in the vertical direction is not smooth and when there is no correlation between the objects on the screen. That is, the first
The vertical change in the field should be gentle, but if the phenomenon in the second field is combined,
This is a case where irregularities are formed in the vertical direction as shown in FIG. This indicates that the object on the screen is moving laterally.

On the other hand, when the expression (1) is satisfied but the expression (2) is not satisfied, the target pixel at the time of the first field scanning and the pixel at the second field scanning which are vertically adjacent to the target pixel. This corresponds to the case where there is a correlation with the pixel. In this case, it is determined that the image is not moving at that point, and the evaluation value C (i, 2j + 1) is set to "0".

Further, the following evaluation formula may be used in order to improve the accuracy of the determination of the moving image / still image. That is, | I (i, 2j) -I (i, 2j + 2) | ≦ δ1 (3) | I (i, 2j) -I (i, 2j + 1) | ≧ δ2 (4) | I (i, 2j + 2) -I (I, 2j + 4) | ≦ δ1 (5) | I (i, 2j + 2) −I (i, 2j + 3) | ≧ δ (6) When all of the expressions (3) to (6) are satisfied, the evaluation value C (I, 2j + 1) is set to "1", and otherwise "0". Then, when ΣC (i, 2j + 1) is δ3 ′ or more, it is determined to be a moving image, and when it is less than δ3 ′, it is determined to be a still image.

By measuring the number of pixels that satisfy the above formulas (1) and (2) or formulas (3) to (6) successively in the x direction, the number of pixels of the object on the screen and the left and right directions You can estimate if you are moving to.

〔The invention's effect〕

As can be easily understood from the above description, according to the present invention, the presence or absence of the low pass filter calculation is selected according to whether the underlying frame signal is a still image or a moving image. Can also form suitable field signals for recording or transmission.

[Brief description of drawings]

FIG. 1 is a flow chart of an embodiment of the present invention, FIG. 2 is a block diagram of a configuration of an apparatus for carrying out the present invention, FIG. 3 is an ideal characteristic diagram of low pass filter calculation, and FIG. 4 is a moving image.・
FIG. 5 and FIG. 6 are configuration block diagrams of the stationary determination device, and FIGS. 5 and 6 are operation explanatory diagrams of the device of FIG. 10 ... Frame memory, 12 ... Digital signal processing circuit, 14 ... Recording / transmission device

Claims (1)

(57) [Claims] 1. A method for extracting a field signal for recording or transmission from a frame signal, comprising determining whether the frame signal accommodated in a frame memory is a moving image or a still image.・ The low-pass filter operation is performed on the frame signal of the memory in the vertical direction, and then the field signal is formed by thinning out pixels in the vertical direction. In the case of a moving image, the low-pass filter operation is not performed. An image processing method, wherein the field signal is formed by thinning out pixels in a vertical direction from a frame signal of a frame memory.