JP2003333370A - Signal processing apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve a problem that a conventional recording apparatus for a video signal adopting the information compression system utilizing the inter- frame correlation typified by the MPEG has difficulty in the compatibility between noise elimination and video resolution in the case of pre-processing of the noise elimination. <P>SOLUTION: A motion vector detected by an information compression circuit of a video signal is fed back to a pre-processing circuit for noise elimination to vary a characteristic of a noise elimination filter in response to motion information of the video signal. Since the visual characteristic is utilized in this case, the noise can be suppressed without recognizing image quality deterioration. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a signal processing device for inputting a video signal from a camera device, a video input terminal or the like and storing the video signal in a recording medium by utilizing a digital image information compression technique, and more particularly, to a signal processing device. It is an object of the present invention to provide a signal processing device including a front signal processing circuit suitable for improving the image quality of a video signal.

[0002]

2. Description of the Related Art Generally, MPEG (Moving) is used as a technique for compressing information of a video signal using time axis correlation data.
There is a technique called "Picture Experts Group", and a video camera using this technique will be described as a conventional technique with reference to FIG.

In the figure, reference numeral 701 is a general circuit block for performing compression signal processing of MPEG technology in a video camera. Reference numeral 201 denotes a video signal input terminal to which an analog video signal is input. Reference numeral 202 is a subtraction circuit that calculates a difference signal between frames of a video signal, and 203 is a quantization circuit that performs DCT (discrete cosine transform) signal processing and quantization processing. Reference numeral 204 denotes an inverse quantization circuit which has the inverse characteristic of the quantization circuit 203 and restores the video signal compressed by the quantization circuit 203 to the original. The output of the inverse quantization circuit 204 is almost the same as the input signal of the quantization circuit 203. Will be equivalent. Reference numeral 205 is an adder circuit for accumulating the output signal of the inverse quantization circuit 204, and 107 is a frame memory for temporarily storing the output video signal of the adder circuit 205 and delaying it for one frame period and outputting it. A frame is a unit representing one screen of a video signal, and is about 33.37 msec in the case of the NTSC system, for example. However, there is a case where one screen is handled by a field (16.68 msec) which is half of one frame by using the interlace technique.

Reference numeral 206 is a motion compensation circuit. The motion compensation circuit 206 is a circuit that detects the correlation between the video signal input from the input terminal 201 and the output signal of the frame memory 107. Since the output signal of the frame memory 107 corresponds to the video signal delayed by one frame, the motion compensation circuit 206 detects the correlation between the current frame image and the frame image one frame before. In the motion compensation circuit 206, the video signal (subject thereof) is
The motion vector data indicating how much and in what direction the motion vector data is calculated, and the output video signal of the frame memory 107 is subjected to the correction corresponding to the motion vector data.

Reference numeral 207 is an encoding circuit for encoding the output video signal of the quantization circuit 203. Entropy coding is used in the MPEG system. 208 is a motion compensation circuit 2
Reference numeral 209 is an encoding circuit for entropy-encoding the motion vector data from 06. Multiplexing 209 multiplexes the video signal encoded by the encoding circuit 207 and the motion vector data encoded by the encoding circuit 208. Circuit.

Reference numeral 210 denotes an output terminal of the multiplexing circuit 209, that is, an output terminal for a compressed video signal compressed by the MPEG system.
A system in which this compressed video signal is converted into a file and recorded on a target medium is generally used. Non-volatile memory cards and optical discs are known as recording media.

The operation of the conventional signal processing apparatus configured as described above will be described below.

First, the video signal input to the input terminal 201 is input to the subtraction circuit 202 and the motion compensation circuit 206. The subtraction circuit 202 subtracts the current frame and the frame one frame before of the input video signal, and outputs the difference signal. The output difference signal is input to the quantization circuit 203, subjected to compression signal processing such as compression (DCT signal processing) and quantization processing by intra-frame correlation, and output to the encoding circuit 207 and the inverse quantization circuit 204. It

The dequantization circuit 204 restores the input compressed signal to the original signal (non-compressed signal), and the addition circuit 205
To perform cumulative addition. The output of the adder 205 is temporarily stored in the frame memory 107 and output to the motion compensation circuit 206 at a timing delayed by one frame.

In the motion compensation circuit 206, one frame of the video signal input to the input terminal and the frame memory 107
The motion vector data indicating how much and in what direction the video has moved during the time of one frame is calculated from the frame delayed by one frame. In addition, the motion compensation circuit 20
6 inputs the frame from the frame memory 107 to the subtraction circuit 202.

The motion vector data from the motion compensation circuit 206 is coded by the coding circuit 208 to be multiplexed by the multiplexing circuit 2.
09 is input.

On the other hand, the compressed video signal from the quantizing circuit 203 is coded by the coding circuit 207 and then multiplexed by the multiplexing circuit 20.
9, the encoded motion vector data is multiplexed and output from the output terminal 210.

As described above, the commonly used MPEG
The video signal compression method of the method is a method of recording a differential signal (that is, a portion different from the previous frame) after adding motion compensation in comparison with the video signal of one frame before.

[0014]

[Problems to be Solved by the Invention]
In the case of the method of recording difference data by utilizing the correlation of the video signal in the time axis direction like the method, the higher the correlation is, the higher the compression rate is, the less the deterioration of the image quality is, and the the lower the correlation is, on the contrary. The signal has a property that the compression rate is low and the image quality is deteriorated. In addition, the video signal that generally exists is
There are few ideal video signals that do not contain noise data, but rather video signals that do contain noise data. When such a video signal is compressed by a method that utilizes correlation, noise data causes a low correlation, which leads to a reduction in compression rate and a deterioration in image quality, which poses a problem.

Therefore, it is known to provide a noise removal processing circuit as a pre-processing before compression. By removing the noise before the compression process (before the input terminal 201 in the configuration of FIG. 7), it is possible to prevent the deterioration of the image quality at the same time without lowering the compression rate. Is performed by band limitation by a low-pass filter, which has a side effect of lowering resolution. Therefore, it was not possible to sufficiently remove noise while maintaining the image quality.

In view of the above problems, the present invention uses the motion vector signal of the video signal to vary the noise removal characteristics based on the human visual characteristics. Therefore, sufficient noise removal is performed before the compression processing. It is an object of the present invention to provide a signal processing device capable of achieving an extremely high image quality as a result.

[0017]

In order to solve the above problems, the present invention provides a first frame memory in a signal processing apparatus for performing signal processing using a frame memory for recording and reproducing a video signal. And a first signal processing circuit using the same and a second frame memory connected to the subsequent stage and a second signal processing circuit using the second frame memory, and detected from the video signal by the second signal processing circuit. The control signal is supplied to the first signal processing circuit and the characteristic of the first signal processing circuit is made variable according to the approximate control signal.

As a result, the noise removal characteristics are made variable based on the human visual characteristics by using the motion vector signal of the video signal, so that sufficient noise removal can be performed before the compression processing, and as a result, It is possible to obtain high image quality.

[0019]

BEST MODE FOR CARRYING OUT THE INVENTION The inventions according to claims 1 to 4 of the present invention include first signal processing means for removing noise from an input video signal, and a video signal from the first signal processing means. A second signal processing means for detecting the movement of the input signal and outputting a control signal and compressing the input video signal, wherein the first signal processing means controls the second signal processing means. It is possible to change the noise removal characteristics based on the signal, and this makes it possible to change the noise removal characteristics based on the human visual characteristics using the motion vector signal of the video signal. Noise can be removed, and as a result, extremely high image quality can be obtained.

Embodiments of the present invention will be described below with reference to the drawings.

(Embodiment 1) FIG. 1 is a block diagram showing a configuration of a signal processing apparatus according to an embodiment of the present invention.

In the figure, 101 is a video input terminal to which an analog video signal is input, and in the digital video camera as in the present embodiment, a video signal captured by the image pickup means is input. Reference numeral 102 is a voice input terminal to which a voice signal is inputted from a sound collecting means such as a microphone, and 103 is a first signal processing means which is a first signal processing means composed of a filter means for removing noise from the inputted video signal. Circuit, 104
Is a first frame memory used for signal processing of the first signal processing circuit 103, and 105 is the first signal processing circuit 10.
3 is a size conversion circuit that is a size conversion unit that converts the image size of the video signal from 3 into a size compatible with MPEG. The size conversion circuit 105 converts the screen size of the input video signal and outputs it. Reference numeral 106 denotes a second signal processing circuit that is a second signal processing unit that performs a compression process on the video signal from the size conversion circuit 105, and outputs the compressed video signal to a multiplexing circuit 109 described later. The calculated motion vector data is fed back to the first signal processing circuit 103 as a control signal. Reference numeral 107 is a second frame memory used for compression processing in the second signal processing means, and 108 is audio signal processing means for performing signal processing such as compression processing on an audio signal input from the audio input terminal 102. A signal processing circuit, 109 is a multiplexing circuit which is a multiplexing means for multiplexing a video signal and an audio signal, and 110 is an output terminal for outputting the video signal and the audio signal output from the multiplexing circuit 109.

In this embodiment, the video signal input from the input terminal 101 is processed by the first signal processing circuit 103 using the frame memory 104, and then the size conversion circuit 105 converts the video signal. Screen size is MPE
After converting into a size corresponding to G (eg, 352 × 288 pixels), the second signal processing circuit 106 performs signal processing using the frame memory 107. At this time, the control signal detected by the second signal processing circuit 106 using the frame memory 107 is fed back to the first signal processing circuit 103 to control the signal processing characteristics. Next, each structure will be described in more detail.

FIG. 2 is a block diagram showing details corresponding to one embodiment of the second signal processing circuit 106 in FIG.

In FIG. 2, 201 is a video signal input terminal to which an analog video signal is input. Reference numeral 202 denotes a subtraction circuit that is a subtraction unit that calculates a difference signal between frames of a video signal, and performs a subtraction process between the frame and the frame one frame before stored in the frame memory 107 to obtain a difference signal (that is, Output different images). 20
A quantizing circuit 3 is a quantizing means for performing intra-frame correlation compression and quantization processing by DCT (discrete cosine transform) signal processing. An inverse quantization circuit 204 is an inverse quantization means that has the inverse characteristic of the quantization circuit 203 and restores the video signal compressed by the quantization circuit 203 to the original. The output of the inverse quantization circuit 204 is the quantization circuit. It is almost equivalent to the input signal of 203. Reference numeral 205 denotes an adding circuit that is an adding unit that accumulates the output signal of the inverse quantization circuit 204.

Reference numeral 107 denotes a frame memory which temporarily stores the output video signal of the adder circuit 205, delays it for one frame period, and outputs it. A frame is a unit representing one screen of a video signal, and in the case of the NTSC system, for example, is about 33.3.
It is 7 msec. However, using the interlace technique, one screen is divided into a field (1
6.68 msec) in some cases.

A motion compensation circuit 206 is a motion compensation means. The motion compensation circuit 206 is a circuit that detects the correlation (also called interframe correlation or temporal correlation) between the video signal input from the input terminal 201 and the output signal of the frame memory 107. Since the output signal of the frame memory 107 corresponds to the video signal delayed by one frame, the motion compensation circuit 206 detects the correlation between the current frame image and the frame image one frame before. This motion compensation circuit 20
In 6, the motion vector data indicating how much and in which direction the video signal (subject thereof) has moved during one frame time is calculated, and the motion vector data is output to the output video signal of the frame memory 107. A process for adding a corresponding correction is performed. The calculated motion vector data is
It is output to the second encoding circuit 208 and the output terminal 211.

Reference numeral 207 is a first encoding circuit which is a first encoding means for encoding the output video signal of the quantization circuit 203. Entropy coding is used in the MPEG system. Reference numeral 208 is a second coding circuit which is a second coding means for entropy-coding the motion vector data from the motion compensation circuit 206. Reference numeral 209 denotes the video signal encoded by the first encoding circuit 207 and the second encoding circuit 20.
8 is a multiplexing circuit that is a multiplexing unit that multiplexes with the motion vector data encoded in 8. Reference numeral 210 denotes an output terminal of the multiplexing circuit 209, that is, an output terminal of a compressed video signal compressed by the MPEG system, and a system for generally recording the compressed video signal into a file and recording it on a target medium. Non-volatile memory cards and optical discs are known as recording media.

These show the general structure of the signal processing of the MPEG system, but are characterized in that a motion vector signal output terminal 211 is newly provided. The motion vector signal is a signal indicating how the video signal moved with respect to the previous frame.

FIG. 3 is a block diagram showing details corresponding to one embodiment of the first signal processing circuit 103.

In the figure, 301 is an input terminal to which a video signal input from the input terminal 101 of FIG. 1 is input, and 302 is a first filter means which is a first filter means for removing horizontal noise in the video signal by band limitation. Directional filter circuit, 3
Reference numeral 03 is a vertical filter circuit which is a second filter means for removing vertical noise in the video signal by band limitation, 304 is a line memory circuit used for noise removal by the vertical filter circuit 303, and 305 is a video In the time direction filter circuit which is the third filter means for removing the noise in the time direction of the signal, the horizontal direction filter circuit 3
02, the vertical direction filter circuit 303, and the time direction filter circuit 305 are composed of band limiting means such as a low-pass filter. 306 is a frame memory that temporarily stores one frame image when noise is removed by the time direction filter circuit 305, and 307 is input to the input terminal 309.
The motion vector signal from the second signal processing circuit 106 is supplied to the horizontal filter circuit 302 and the vertical filter circuit 30.
3 and the time direction filter circuit 305, a discrimination circuit which is a discrimination means for discriminating, and 308 is each filter circuit 302 and 30
The output terminal of the video signal from which noise has been removed by 3 and 305 is output to the size conversion circuit 105 of FIG. Three
Reference numeral 09 is an input terminal for a motion vector signal from the second signal processing circuit 106.

The horizontal filter circuit 302 limits the horizontal band of the video signal. Therefore, if the band is strongly restricted, the noise component is suppressed and the image quality is improved, but the resolution in the horizontal direction is deteriorated. The vertical filter circuit 303 also uses the line memory 304 to limit the vertical band of the video signal. Also in this case, if the band is strongly restricted, the image quality is improved, but the resolution in the vertical direction is deteriorated. Further, the time direction filter circuit 305 uses the frame memory 306 to limit the band in the time direction. The band limitation in the time direction does not cause the problem of resolution deterioration, but if the band is strongly limited, there is a problem that an afterimage phenomenon occurs in exchange for the improvement of image quality.

In this embodiment, the characteristics of the horizontal filter circuit 302, the vertical filter circuit 303, and the temporal filter circuit 305 are input from the input terminal 309 based on the human visual characteristics. , And is controlled according to the motion vector signal from the second signal processing circuit 106. An example of this control will be described with reference to FIG.

FIG. 6 is a schematic diagram for explaining the motion vector signal and the filter operation controlled by the motion vector signal. The motion compensation circuit 2 in the second signal processing circuit 106 is shown in FIG.
Based on this schematic diagram, how the above filter circuit is operated is controlled by the motion vector detected in 06.

In the figure, the vertical axis represents the amount of movement of the video signal in the vertical direction. That is, when the value of the motion vector signal is in the central part, it indicates that the video signal has no vertical motion with respect to the previous frame, and when the value of the motion vector signal is in the upper part, the video signal is Indicates that motion is occurring in the positive direction in the vertical direction, and that the value of the motion vector signal is lower indicates that motion is occurring in the negative direction in the vertical direction in the video signal.

The horizontal axis indicates the amount of horizontal movement of the video signal. That is, when the value of the motion vector signal is in the central part, it means that the video signal has no horizontal motion with respect to the previous frame, and when the value of the motion vector signal is in the right part, This indicates that the signal is moving in the horizontal positive direction, and the case where the value of the motion vector signal is the left part indicates that the video signal is moving in the horizontal negative direction.

In the figure, the area indicated by is the area of the motion vector when there is no motion in the vertical and horizontal directions of the video signal, that is, the video is almost stationary. It is known that human visual sensitivity is extremely high for a video signal that is not moving. Therefore, in the area where the motion vector is, the characteristics of the horizontal direction filter 302 and the vertical direction filter 303 are set to OFF or weak, and the characteristics of the time direction filter 305 are set to strong. By doing so, the afterimage phenomenon does not occur even if the filter characteristic in the time direction is strongly set for the video signal that is not moving.

The area shown below is an area in which the video signal has a vertical motion but no horizontal motion. In this case, the characteristics of the vertical direction filter 303 are set strongly, and the characteristics of the horizontal direction filter 302 and the time direction filter 305 are set to OFF or weak.
Even if the resolution in the vertical direction deteriorates in the video signal moving in the vertical direction, it is difficult to visually confirm, and
It is difficult to recognize that the image quality is deteriorated.

Further, the area indicated by is an area where the video signal has a horizontal motion, but does not have a vertical motion. In this case, the horizontal filter 302
Is set to be strong, and the characteristics of the filter 303 in the vertical direction and the filter 305 in the time direction are set to OFF or weak. Even if the horizontal resolution of a video signal moving in the horizontal direction is degraded, it is difficult to visually confirm it, and it is difficult for the user to recognize that the image quality is degraded.

The last area is an area where motion in the licking direction occurs in the video signal. In this case, the characteristics of the horizontal filter 302 and the vertical filter 303 are set strong, and the characteristics of the time filter 305 are set weak.

By controlling the characteristics of each filter according to the motion vector signal in this way, it is possible to suppress the noise component without visually recognizing deterioration of the image quality by the user.

The division of the area described with reference to FIG. 6 is merely an example of the embodiment, and the feature of the present invention is that the various filter characteristics are switched according to the motion vector signal.

Next, a specific embodiment of the size conversion circuit 105 will be described with reference to FIG. In the figure, reference numeral 401 denotes a video signal input terminal to which the video signal from the first signal processing circuit 103 is input. Reference numeral 402 denotes a horizontal resizing circuit that changes the horizontal image size (pixels) of the input video signal,
403 is the vertical image size of the video signal (pixels)
Vertical line resize circuit 404 for changing the line size, 404 a line memory for temporarily storing a line of a video signal at the time of resize in the vertical direction resize circuit 403, 405 a register circuit for preliminarily storing a predetermined image size, 40
Reference numeral 6 is an output terminal for the resized video signal, and the output video signal is input to the second signal processing circuit 106.

According to the value set in the register circuit 405, the video signal is first resized in the horizontal direction by the horizontal resize circuit 402, and then resized in the vertical direction by the vertical resize circuit 403. Generally, the video signal is resized from an image having a large number of pixels to an image having a smaller number of pixels.

Further, as shown in the figure, the line memory 404 is required for the resize signal processing in the vertical direction.
Therefore, by performing the horizontal resize signal processing first to reduce the data amount of the video signal and then performing the vertical resize signal processing, the number of stages of the line memory 404 can be reduced and the circuit scale can be reduced. Have contributed.

A concrete example of the horizontal and vertical resize signal processing is shown in FIG. FIG. 5A is a schematic diagram showing a part of the input video signal, showing pixels in block units. FIG. 7B shows a video signal that has been resized in the horizontal direction, and FIG. 7C shows a video signal that has been resized in the vertical direction. As shown in FIG. 5A, a part of the input video signal is input in
The output signal of the horizontal resizing circuit is represented by (m, n), and the output signal of the resizing circuit 105 is represented by A (m, n).
When expressed by ut (m, n), the calculation in the signal processing is expressed as shown in FIG. However, this is merely an example, and the present invention is not limited to this.

In FIG. 5, the input video signal shown in FIG. 5A is first resized in the horizontal direction by the horizontal resizing circuit 402. Specifically, in each pixel in FIG. 5A, two horizontally adjacent pixels are combined,
As shown in FIG. Next, in the vertical direction resizing circuit 403, two pixels adjacent in the vertical direction in each pixel of FIG. 5B are combined to obtain the structure shown in FIG. 5C. As a result, the number of pixels output from the vertical resizing circuit 403 can be halved in the horizontal and vertical directions as compared with the pixels of the input video signal. In addition,
In order to make the explanation easy to understand in the drawings, FIG.
The size of each pixel in (c) is shown in FIGS.
Although it is illustrated that the size of the image becomes larger, the size of the pixel does not actually change, and the size of the entire image (vertical and horizontal dimensions) is reduced by reducing the number of pixels.

As described above, according to this embodiment, the second
Before the compression processing in the signal processing circuit 106, the first signal processing circuit 10 based on the motion vector signal of the video signal.
By performing noise removal in step 3, it is possible to control the movement in the direction in which the deterioration of resolution is less noticeable by human vision by applying a stronger filter in a range where the deterioration of resolution is not noticeable. It is possible to remove various noises and improve the image quality.

[0049]

As described above, according to the present invention, for example, M
When compressing and recording a video signal using a method of recording difference data using the correlation of the video signal in the time axis direction, such as the PEG system, noise data causes a reduction in compression rate and deterioration in image quality. However, in the present invention, since the noise removal characteristics are made variable based on the human visual characteristics using the motion vector signal of the video signal, sufficient noise removal can be performed before the compression processing, and as a result, It is possible to obtain high image quality.

[Brief description of drawings]

FIG. 1 is a block diagram of a signal processing device according to an embodiment of the present invention.

FIG. 2 is a block diagram of a second signal processing circuit according to an embodiment of the present invention.

FIG. 3 is a block diagram of a first signal processing circuit according to an embodiment of the present invention.

FIG. 4 is a block diagram of a resizing circuit that is an embodiment of the present invention.

FIG. 5 is a schematic diagram for explaining the operation of the resizing circuit according to the embodiment of the present invention.

FIG. 6 is a schematic diagram for explaining the operation of the motion vector signal discriminating circuit according to the embodiment of the present invention.

FIG. 7 is a block diagram of a compression circuit of a conventional signal processing device.

[Explanation of symbols]

101 Video signal input terminal 102 Audio signal input terminal 103 first signal processing circuit 104 first frame memory 105 size conversion circuit 106 Second signal processing circuit 107 Second frame memory 108 audio signal processing circuit 109 Multiplexing circuit 110 output terminals

Claims (4)

[Claims] 1. A first signal processing means for removing noise from an input video signal, and an input video which detects a movement of the video signal from the first signal processing means and outputs a control signal. A second signal processing means for compressing a signal, wherein the first signal processing means is capable of varying noise removal characteristics based on a control signal from the second signal processing means. Processing equipment. 2. The control signal is vector data indicating a correlation of a video signal in the time direction.
The signal processing device described. 3. The signal processing device according to claim 1, wherein the first signal processing means is a signal processing circuit having a noise removing function of a video signal. 4. The signal processing apparatus according to claim 3, wherein the first signal processing means includes a horizontal filter circuit, a vertical filter circuit, and a time filter circuit.