JP3537189B2 - Television signal converter - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention
The present invention relates to a television signal converter that converts numbers and removes distortion due to the conversion by interpolation using a motion vector signal, and is applicable to, for example, a television system converter.

[0002]

2. Description of the Related Art In a television system converter,
It is necessary to convert the number of horizontal scanning lines (the number of lines) and the number of fields. Generally, these conversions are performed by digitizing a television signal, storing it in a memory having a capacity of two or more fields, and controlling the address at the time of reading, and further interpolating the read television signal by a line. Processing and field interpolation processing are performed to remove distortion due to conversion.

Here, the field interpolation processing includes:
(1) Field interpolation by two-field linear interpolation, (2)
Field interpolation using four fields, (3) field interpolation using motion vectors, etc.
Journal of the Institute of Television Engineers of Japan vol. 45, no. 12, p
p. 1534-1543: Note that this document describes the entirety of television system conversion in detail). Of these, field interpolation using motion vectors corrects by capturing the moving image itself, so that a natural moving image without degradation in resolution and without jerkiness can be obtained. Is applied to many television format converters which are intended to improve image quality.

FIG. 2 shows an example of the configuration of a conventional television system conversion apparatus adopting such field interpolation using a motion vector (for example, Reference 2).
See JP-A-1-309597. Note that input / output signals of the television system converter are component signals (digital signals) of the luminance signal Y, the red color difference signal RY, and the blue color difference signal BY.

In FIG. 2, a luminance signal Y input from an input terminal 21a is input to a motion detecting circuit 23, which obtains a motion detecting signal from a difference signal between one frame. Is input to the three-field memory 25 and stored. Also, component signals Y from the respective input terminals 21a, 21b, 21c,
RY and BY are delayed by the processing delay in the motion detection circuit 23 via the delay circuit 24, input to the three-field memory 25, and stored.

The conversion of the number of lines and the number of fields are performed through reading control from the three-field memory 25. The two-field luminance signals simultaneously read from the three-field memory 25 are input to the intra-field line interpolation circuits 26Y1 and 26Y2, respectively, subjected to intra-field line interpolation processing, and switched to the switching circuit 28.
Y, and two-field luminance signals read simultaneously from the three-field memory 25 are input to the inter-field line interpolation circuit 27Y, subjected to inter-field line interpolation processing, and input to the switching circuit 28Y. You.
The switching circuit 28Y includes a three-field memory 25
The motion detection signal read out in synchronization with the readout of the luminance signal of two fields is input to the switching circuit 28Y.
Selects an inter-field line interpolation signal for a still image and a field interpolation signal for a moving image based on a motion detection signal, and suppresses degradation in resolution due to line interpolation.

For a color signal (RY or BY),
In-field line interpolation circuits 26C1, 26C2, inter-field line interpolation circuit 27C, and switching circuit 28C
Performs the same processing. The switching circuit 2
8C, the same motion detection signal is input to the switching circuit 28Y.

The two-field luminance signals for which the line interpolation method has been selected by the switching circuit 28Y are supplied to a motion vector detection circuit 29 and a field interpolation circuit 31Y. The two-field color signals for which the line interpolation method has been selected by the switching circuit 28C are supplied to the field interpolation circuit 31C.

The motion vector detection circuit 29 divides the image into blocks of 8 pixels × 8 lines, and detects a motion vector using a signal separated by one field for each block, for example, according to a repetitive gradient method. The motion vector signal is supplied to the motion vector selection circuit 30.

The motion vector selection circuit 30 evaluates the detected motion vector based on the interpolation field,
When it is determined that the motion vector signal is valid, the motion vector signal is supplied to both the field interpolation circuits 31Y and 31C to perform field interpolation using the motion vector signal, and when it is determined that the motion vector signal is invalid, Supplies a signal instructing a time-axis linear interpolation process (weighted addition of two-field signals) to both of the field interpolation circuits 31Y and 31C to execute the field interpolation according to the time-axis linear interpolation method. In this manner, the discontinuity of the moving image caused by the conversion of the number of fields is removed by field interpolation.

As described above, the luminance signal Y, red color difference signal RY and blue color difference signal BY subjected to the conversion of the number of lines and the number of fields, line interpolation and field interpolation are output to the output terminal 22a, respectively. , 22b, 22c.

[0012]

However, the conventional television system converter has the following problems.

(1) In the case of a signal having a small afterimage effect such as a CCD camera, the correlation between fields becomes small, and a detection error of a motion vector easily occurs. In order to correct this, it is known to detect a motion vector after filtering a signal for detecting a motion vector in a time axis direction. However, inserting a time axis filter into a discontinuous signal after the number of fields has been converted causes a significant increase in the circuit configuration.

(2) Since the detection of the motion vector is performed after the line interpolation, the line interpolation processing and the field interpolation processing are separated (simultaneous processing cannot be performed). Therefore, the image quality is deteriorated as compared with the simultaneous processing.

(3) Since the detection of the motion vector is performed after the line interpolation, the accuracy of the motion detection used for the line interpolation affects the detection accuracy of the motion vector. This is because the inter-field difference between two fields separated by one field and subjected to intra-field line interpolation processing becomes large for an image having a small field correlation, and the inter-field difference value used for motion vector detection becomes the motion amount. This is because it will not respond. For example, a motion vector detection error is likely to occur for a television signal having a small horizontal level change but a large vertical level change, such as a television signal obtained by imaging a blind.

(4) Since the detection of the motion vector, which is the processing for each block, is performed after the line interpolation, two-field signals separated by one field need to be completed. As described above, when the synchronization on the input side and the synchronization on the output side are independent of each other, a memory having a three-field capacity is required as described above in consideration of a buffer for preventing temporal jump of a field. That is, the required memory capacity is considerably large.

There is a television signal converter such as a television converter which only converts the number of fields or frames without converting the number of lines.
Even in such a television signal conversion device, some of the problems described above are directly considered as problems.

[0018]

In order to solve the above-mentioned problem, in the present invention, at least the number of fields is converted through writing and reading to and from an image memory. In a television signal converter for removing by an insertion process, a television signal input to an image memory is branched and input, and a first field of the television signal is input.
Of the two lines, and the signal located in the middle of the two lines
Based on the signal of one line of two fields, the second field
Form a signal located at the same position as the
A pre-filter means for deinterlacing the lines of the first field and the second field, and forming a motion vector signal based on the television signal output from the pre-filter means to form an image memory And the image memory writes the obtained motion vector signal together with the television signal before the number of fields is converted, and the television after the number of fields is converted. The reading is performed in synchronization with the John signal.

[0019] Here, if having a line number conversion configuration and the number of lines interpolation configuration, shared with the image memory for the number of lines transform, and an image memory fields Variable changeover motion
A vector detection circuit is installed on the input side of this shared image memory.
Line interpolation and line position in each field
Line interpolation circuit that performs non-interlacing to align the positions
Is provided on the output side of the shared image memory, and a line interpolation circuit is provided.
In the subsequent stage, the above motion vector read from the shared image memory
In the field where field interpolation is performed using vector signals
Preferably, an insertion circuit is provided .

[0020]

In the present invention, the television signal before the number of fields is converted, more specifically, two lines of the first field of the television signal input to the image memory
Signal and the second field located between the two lines
Of the second field based on the signal of one line of the second field.
A signal which is located at the same position as the input signal is formed, and a field interpolation process is performed based on the television signal before the number of fields in which the lines of the first field and the second field are aligned is converted. A motion vector signal to be used is formed. Therefore, compared with the conventional method of performing motion vector detection after the line interpolation circuit on the output side,
The noise of the input signal to the motion vector detection circuit can be removed, stable and accurate motion vector detection can be performed, and the overall configuration can be reduced in size. Further, in the present invention, the motion vector detection circuit detects a motion vector in synchronization with the input television signal, and the image memory converts the obtained motion vector signal into a television signal in which the number of fields is converted. The writing is performed together in synchronization with the signal, and the reading is performed in synchronization with the television signal whose number of fields is converted. This eliminates the need for a buffering period of the input television signal in consideration of the detection of the motion vector. Thus, the capacity of the image memory can be reduced.

Further, in the case of having the line number conversion configuration and the line number interpolation configuration, the image memory for converting the number of lines and the image memory for converting the number of fields or the number of frames are used in common. If a motion vector signal is formed based on a television signal on the input side of the shared image memory, non-interlacing required for field interpolation can be provided as a function of a line interpolation configuration, and the Advantages such as reduction of hardware can be achieved.

[0022]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment in which the present invention is applied to a television converter will be described in detail with reference to the drawings. Here, FIG. 1 is a block diagram showing the overall configuration of this embodiment.

In FIG. 1, the television system converter of this embodiment includes a delay circuit 1, a pre-filter 2, a one-field delay circuit 3, a motion vector detection circuit 4, a two-field memory 5 for converting the number of lines and the number of fields, It comprises a line interpolation circuit 6 and a field interpolation circuit 7.

The luminance signal (digital signal) input from the input terminal 8 is supplied to the delay circuit 1 and the pre-filter 2.

The pre-filter 2 is a spatio-temporal filter provided to improve the accuracy of detecting a motion vector based on a two-field signal, and performs band limitation and non-interlacing (correction of the center of gravity) of an input luminance signal. is there.

FIG. 3 shows the position of the center of gravity of the input signal and the output signal of the pre-filter 2. In order to detect a motion vector between fields, it is necessary to obtain an inter-field difference. For that purpose, it is assumed that a signal exists at the same position in both fields. However, since the television signal (the luminance signal in this embodiment) is interlaced, no signal exists at the same position between the two fields. Therefore, it is necessary to Non'inta race of which such signals are present at the same position of both field before performing the detection of the motion vector. In the case of this embodiment, as shown in FIG.
The pre-filter 2 has a first field (odd field)
2 line signal and 2nd field (even field)
Weighted addition using weighting coefficients of 1/4, 1/4, and 1/2 is performed on a signal of a total of three lines including a signal of one line located in the middle of the second line and a line of the second field. A signal located at the same position is formed and output. Incidentally, the number of lines of the input signal and the output signal is Non'in sauce over gasification here is the same.

When a motion vector is detected from a line interpolation signal, a pre-filter may be provided before the field interpolation circuit to increase the detection accuracy.
Such prefilter, because the conversion of the number of interpolation and field lines directed to the already made signal, required to be appropriately changing the weighting factor or the like to perform Non'inta race of, as a result, the configuration is fairly Large and complicated.

The output signal from the pre-filter 2 is delayed by one field via a one-field delay circuit 3 and given to a motion vector detection circuit 4 as a previous field signal 9, and supplied to the motion vector detection circuit 4 by a current field signal. Provided directly as signal 10. That is, two fields of signals are input to the motion vector detection circuit 4.

The motion vector detecting circuit 4 may use any method. For example, the motion vector detecting circuit 4 detects a motion vector according to a repetitive gradient method and stores the motion vector signal 11 in the two-field memory 5 for converting the number of lines and the number of fields. give.

FIG. 4 shows an example of the configuration of the motion vector detection circuit 4 (see the above-mentioned document 1). In FIG. 4, the initial deviation vector selection unit 41 performs block matching of the previous field signal 9 and the current field signal 10 from the detected motion vectors near the detected block stored in the motion vector memory 40. An initial displacement vector Vo is selected based on this. A vector Vp corresponding to a deviation from the initial deviation vector Vo is obtained by a gradient method operation by the first gradient method operation unit 42 using the initial deviation vector Vo, the previous field signal 9 and the current field signal 10, and added. The minute vector is corrected by the unit 43. Further, the vector Vo + Vp, the vector Vo by the gradient method operation by the second gradient method operation unit 44 using the previous field signal 9 and the current field signal 10.
A vector Vq corresponding to a deviation from + Vp is obtained, and the vector is corrected by the adder 45. The corrected vector is output as a final motion vector V (= Vo + Vp + Vq) and is also provided to the motion vector memory 40 and stored. .

The motion vector detection circuit 4 may convert a motion vector signal obtained in units of blocks into a motion vector signal according to a raster scan in the same manner as the input luminance signal, and output the signal. In this way, the two-field memory 5 can easily access the motion vector signal.

On the other hand, the luminance signal input from the input terminal 8 is delayed by the delay circuit 1,
A memory input luminance signal 12 time-aligned with 1 is input to the two-field memory 5.

The two-field memory 5 for converting the number of lines and the number of fields is a buffer memory for converting the number of lines and the number of fields of an input luminance signal, and depends on the television system on the input side and the output side. For a luminance signal, it has a capacity of at least two fields,
Further, the motion vector signal only needs to have a capacity of at least one field corresponding to the current field luminance signal. The block showing the two-field memory 5 in FIG. 1 also includes a write control circuit and a read control circuit of the memory.

Writing of the luminance signal to the two-field memory 5 is performed in synchronization with the input luminance signal, and reading is performed in synchronization with the output side to be converted. Here, reading from the memory 5 is performed simultaneously for two fields, that is, the previous field luminance signal 13 and the current field luminance signal 14 are simultaneously output. The writing of the motion vector signal to the two-field memory 5 is also performed in synchronization with the input luminance signal, and the reading from the memory 5 is also performed in synchronization with the output side to be converted.

Through the reading of the luminance signal, the number of lines and the number of fields are converted. For example, the television system converter according to the present embodiment changes the NTSC system based on an interlace signal of 60 fields per second and 262.5 lines per field from the PAL system based on an interlace signal of 50 fields per second and 312.5 lines per field. For conversion of the system, reading is performed according to the PAL system, and the number of lines and the number of fields are converted.

The previous field luminance signal 13 and the current field luminance signal 14 simultaneously read from the two-field memory 5 are input to the line interpolation circuit 6. The line interpolation circuit 6 removes distortion due to the line number conversion through the line interpolation processing, performs non-interlacing for the field interpolation processing, and outputs the previous field luminance signal after the line interpolation. 16 and current field luminance signal 17
To the field interpolation circuit 7.

The field interpolation circuit 7 also receives the motion vector signal 15 read from the two-field memory 5. That is, the motion vector signal 15 related to the previous field luminance signal 16 and the current field luminance signal 17 from the line interpolation circuit 6 is input. Note that a delay circuit for a motion vector signal that compensates for a delay caused by the line interpolation processing may be built in. The field interpolation circuit 7 performs motion compensation using the motion vector signal 15 and then performs field interpolation using the previous field luminance signal 16 and the current field luminance signal 17 to suppress the image quality deterioration due to the field number conversion. Output terminal 1 by forming a final luminance signal according to the
8 to the next stage circuit.

Conventionally, since the field interpolation circuit uses a motion vector signal, it is based on the idea that a motion vector signal is formed from a signal input to the field interpolation circuit. However, it has the disadvantages described above. Therefore, in this embodiment, the field interpolation circuit 7 uses a motion vector signal detected by a signal before being input to the line number field number conversion two field memory 5.

The field interpolation circuit 7 in this embodiment
May have any configuration as long as it has a configuration for performing motion correction using the motion vector signal 15.

FIG. 5 shows a detailed configuration example of the field interpolation circuit 7 (see the above-mentioned document 1). In FIG. 5, the previous field luminance signal 16 and the current field luminance signal 17 are motion-corrected by motion vector amounts αV and (1−α) V corresponding to the field interpolation ratio α and (1−α), respectively, and are stored in the memory 50. , 51 and then read from these memories 50 and 51 and read out from the two multipliers 52 and 5
3 and the field interpolation ratio α, (1
-Α) is added. The previous field luminance signal 16 and the current field luminance signal 17 are temporarily stored in the corresponding memories 55 and 56, respectively, read from the memories 55 and 56, and read out from the memories 55 and 56 by two multipliers 57 and 58 and an adder. 59, weighted addition is performed according to the field interpolation ratio α, (1−α). In this way, a field interpolation signal with motion compensation using the motion vector V (15) is output from the adder 54 ,
The adder 59 outputs a linearly interpolated field interpolation signal. These two field interpolation signals are the switching coefficients β, (1) output from the adaptive switching coefficient generation circuit 60.
−β), weighted addition is performed by the two multipliers 61 and 62 and the adder 63, and is output as a final field interpolation signal.

Hereinafter, the operation of the television system conversion apparatus according to the embodiment including the above-described units will be described.

The main function of the television system converter is to convert the number of lines and the number of fields. The conversion of the number of fields is made by interpolating a new field from a signal of two consecutive fields. At this time, if the motion vector between the two fields is detected, and the motion is corrected and interpolated, the quality of the converted image can be greatly improved. The television system converter of this embodiment operates to effectively exhibit such a function.

The luminance signal (digital signal) input from the input terminal 8 is supplied to the delay circuit 1 and the pre-filter 2.

The luminance signal is band-limited and non-interlaced (correction of the center of gravity) by the pre-filter 2, and is supplied to the motion vector detection circuit 4 as the previous field signal 9 via the one-field delay circuit 3.
The current field signal 10 is directly supplied to the motion vector detection circuit 4. Thus, a motion vector is detected by the motion vector detection circuit 4 and the obtained motion vector signal 11 is supplied to the two-field memory 5 for line number and field number conversion in synchronization with the input luminance signal 12 via the delay circuit 1. Stored.

Through reading of the previous field luminance signal 13 and the current field luminance signal 14 from the two-field memory 5, conversion to the number of lines and conversion of the number of fields according to the television system on the output side are executed, and in synchronization with the reading. Thus, the motion vector signal 15 related to the previous field luminance signal 13 and the current field luminance signal 14 is also read.

The read previous field luminance signal 13 and current field luminance signal 14 are line-interpolated by the line interpolation circuit 6, and through this processing, distortion due to line number conversion is removed, and the next field interpolation is performed. Is deinterlaced in consideration of

The previous field luminance signal 16 after line interpolation
The current field luminance signal 17 is subjected to motion correction using the motion vector signal 15 by the field interpolation circuit 7 and then weighted and added (field interpolation processing is executed), and output as a final converted output signal It is sent from terminal 18.

According to the above embodiment, the motion vector detection circuit 4 for detecting the motion vector signal used by the field interpolation circuit 7 detects the motion vector signal from the input television signal whose number of lines and the number of fields have not been converted. As a result, the three-dimensional pre-filter 2 can be inserted before the motion vector detection circuit 4, and compared with the conventional method of performing motion vector detection after the output-side line interpolation circuit,
The noise of the input signal to the motion vector detection circuit can be removed, stable and accurate motion vector detection can be performed, and the overall configuration can be reduced in size.

Further, according to the above embodiment, since the motion vector signal is synchronized with the input side and the output side, the buffering period of the input television signal in consideration of the detection of the motion vector becomes unnecessary, and the number of fields is reduced. Two fields are sufficient for the capacity of the conversion field memory (5).

Further, according to the above-described embodiment, a necessary non-interlace circuit can be provided as a function in the line interpolation circuit 6 in a stage preceding the field interpolation circuit 7, so that the hardware can be greatly reduced. Since no motion vector detecting circuit exists between the line interpolation circuit 6 and the field interpolation circuit 7, the line interpolation circuit 6
In addition, simultaneous processing in a form in which the field interpolation circuit 7 is integrated can be performed, and the image quality can be improved.

In the above embodiment, the luminance signal Y
However, the present invention can be applied to a television system converter for a red color difference signal RY or a blue color difference signal BY, and further, primary color signals R, G, The present invention can be applied to a television system converter for B. Further, similarly to the above-described conventional example, the motion vector signal obtained for the luminance signal Y is
The color difference signals RY and BY may be used in a conversion system.

In the above embodiment, the present invention is applied to a television system converter, but may be applied to other television signal converters. For example, the present invention can be applied to so-called telecine for exposing a television signal to a cinema film or so-called kinetere for converting an image on the cinema film into a television signal. Therefore, the number of unit images to be converted is not limited to a field, but may be a frame.

[0053]

In the above embodiment, the motion vector signal 15 is used by the field interpolation circuit 7. However, the line interpolation circuit 6 may be used as a motion signal.

[0055]

As described above, according to the present invention, the motion vector detecting circuit for forming the motion vector signal detects the motion vector in synchronization with the input television signal before the number of fields is converted. And the image memory writes the obtained motion vector signal together with the television signal whose field number is converted, and reads it out in synchronization with the television signal whose field number is converted. Therefore, it is possible to realize a small-sized television signal converter capable of improving the image quality of the converted television signal.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of an embodiment.

FIG. 2 is a block diagram illustrating a configuration of a conventional television format conversion device.

FIG. 3 is an explanatory diagram showing an input / output relationship of the pre-filter of the embodiment.

FIG. 4 is a block diagram illustrating a configuration example of a motion vector detection circuit 4 according to the embodiment.

FIG. 5 is a block diagram illustrating a configuration example of a field interpolation circuit 7 according to the embodiment.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Delay circuit, 2 ... Pre-filter, 3 ... 1-field delay circuit, 4 ... Motion vector detection circuit, 5 ... Line number and field number conversion 2 field memory, 6 ... Line interpolation circuit, 7 ... Field interpolation circuit .

──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Takushi Iwamoto 2-2-1 Jinnan, Shibuya-ku, Tokyo Japan Broadcasting Corporation Broadcasting Center (72) Inventor Kenji Nakajima 2-2-1 Jinnan, Shibuya-ku, Tokyo Japan Within the Broadcasting Corporation Broadcasting Center (72) Jun Yamakita Inventor 2-2-1 Jinnan, Shibuya-ku, Tokyo Japan Broadcasting Corporation Broadcasting Center (56) References JP-A-1-309597 (JP, A) JP-A-4- 372292 (JP, A) JP-A-4-78284 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) H04N ^7/ 00-7/088

Claims (2)

(57) [Claims] 1. A television signal conversion apparatus for converting at least the number of fields through writing and reading to and from an image memory and removing distortion caused by the conversion by a line interpolation process or a field interpolation process using a motion vector signal. The television signal input to the image memory is branched and input, and the first field of the television signal is input.
Of the two lines, and the signal located in the middle of the two lines
Based on the signal of one line of two fields, the second field
Form a signal located at the same position as the
A non-interlaced pre-filter for aligning the positions of the lines of the first field and the second field; and forming the motion vector signal based on the television signal output from the pre-filter. A motion vector detection circuit to be provided to the image memory, and the image memory writes the obtained motion vector signal together with the television signal before the field number is converted, and the field number is converted. A television signal conversion device for reading in synchronization with a subsequent television signal. 2. The television signal conversion device according to claim 1, wherein the television signal conversion device has a line number conversion configuration and a line interpolation configuration, and shares an image memory for converting the number of lines and an image memory for converting the number of fields. The motion vector detection circuit is provided on the input side of the shared image memory, and the line interpolation circuit for performing line interpolation and non-interlacing for aligning line positions in each field is provided on the output side of the shared image memory. A television which is provided at a subsequent stage of the line interpolation circuit and performs a field interpolation using the motion vector signal read from the shared image memory. Signal converter.