JP2837925B2 - Motion vector detection method of TV converter - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a system for detecting motion information of a picture, that is, a motion vector in a TV system conversion device using a digitalized TV (television) signal. .

(Conventional technology) The technology for performing motion compensation using the above-described motion vector is used to improve the inter-frame coding efficiency in high-efficiency coding of a TV signal, or to perform TV system conversion (conversion between NTSC system and PAL system). , Etc.) to reduce the discontinuity of motion due to the conversion of the field number.

For the detection of this motion vector, the TV signal is
There is a method of detecting a motion vector for each block after subdividing into blocks of Xn lines (m and n are integers).
The pattern matching method disclosed in Japanese Patent Application Laid-Open Nos. 162683 and 162684 and the iterative gradient method disclosed in Japanese Patent Application Laid-Open No. 60-158786 are well known.

FIG. 3 shows an example of a conventional motion vector detecting circuit used in a TV system converter. (For reference, proceedings of the 1989 National Convention of the Institute of Television Engineers of Japan, 20-
5, P501-502) The detection of the motion vector is performed by dividing the pixels into blocks of 8 pixels × 8 lines and for each block. The input signal 41 of the same figure inputs the current field luminance signal, and the previous field luminance signal 42 inputs the two-dimensional luminance signal.
The LPF (Low Pass Filter) removes noise and high frequency components. The removed signal is input to an initial displacement vector selection circuit 46 and a displacement vector detection circuit 47.
A motion vector memory 45 stores the detected motion vector.

The initial displacement vector selection circuit 46 reads several types of detected motion vectors in a block near the detected block from the motion vector memory circuit 45, and selects an optimum motion vector from these. The optimal determination is made as follows. That is, the coordinates of the block are displaced by the amount of each motion vector, the absolute value of the difference value between the fields is obtained, and the motion vector that gives the block having the smallest value obtained by accumulating the values in the block is determined to be optimal. The motion vector thus selected is supplied to a displacement vector detection circuit 47. The detection circuit 47 detects the vector by using the iterative gradient method after deviating the coordinates of the block corresponding to the initial displacement vector. . The displacement vector and the above-mentioned initial displacement vector are added by an adding circuit 48 to obtain a motion vector to be obtained.
Then, the motion vector is stored in the motion vector memory circuit 45.

(Problems to be Solved by the Invention) In any of the conventional methods described above, since the motion vector is detected using only the luminance signal, the level change of the luminance signal and the level change of the color difference signal even without motion are detected. When there is a motion or a motion, the motion vector is not detected. Therefore, image distortion occurs in the motion interpolation image using the motion vector. For example, in the case of an image as shown in FIG. 4, that is, when the luminance signal is a signal of a constant level, while the color difference signal (the RY signal in the figure) is a window signal moving as shown in FIG. Since the vector is 0, this motion is not detected.

The present invention solves this detection of insufficient motion vectors,
An object of the present invention is to provide a method capable of detecting a motion vector faithful to a motion.

(Means for Solving the Invention) In order to solve the above-mentioned problems, the present invention separately detects a motion vector for a luminance signal and a color difference signal, and calculates a luminance signal if an image gradient calculation result of the luminance signal is equal to or larger than a threshold value. The motion vector detected by the signal is used as the desired motion vector,
When the result of the image gradient calculation is smaller than the threshold value, the motion vector detecting the color difference signal is used as the motion vector to be obtained. When there are a plurality of types of color difference signals, generally two types (RY, BY), when the color difference signal is selected below the threshold value, the larger one of the two types is selected.

(Operation) Since the present invention employs the method described above, it is possible to detect a motion vector that is faithful to the motion of an image. That is, even when there is no level change or motion of the luminance signal and there is a level change or motion of the color difference signal, the motion vector is detected, and the image distortion is sufficiently eliminated.

(Embodiment) FIG. 1 shows a motion vector circuit diagram of a first embodiment of the present invention. This figure shows a method of detecting a motion vector in a TV system conversion apparatus by subdividing a TV signal into blocks of 8 pixels × 8 lines, as described in the section of the related art, and then detecting each block.

1 is the luminance signal of the current field, 2 is the luminance signal of the previous field, 3 is the chrominance signal RY of the current field, 4 is the RY of the previous field, 5 is BY of the current field, 6 is the luminance of the previous field. BY, and these signals are input. The input signals are subjected to two-dimensional LPFs (low-pass filters) 7 to 12 to remove noise and reduce high-frequency components of the image, and correspond to initial deviation vector selection circuits 14 to 16 and gradient method operation circuits 17 respectively. Enter in ~ 19. Also,
The output of the two-dimensional LPF 7 of the luminance signal 1 of the current field is also input to the image gradient calculation circuit 27.

A motion vector memory 13 stores the detected motion vector 40 in the same manner as described in the section of the related art.

The initial deviation vector selection circuits 14 to 16 are also circuits similar to the functions described in the section of the related art, and read out several types of detected motion vectors in blocks near the detected block from the memory 13 and optimize the Out what is right. The displacement vector detection circuits 17 to 19 are also circuits having a function of calculating by the iterative gradient method, similarly to the conventional one. Both circuits 14 and 17, 15 and 18,
The outputs of 16 and 19 are added by adding circuits 20 to 22 to obtain motion vectors 23 to 25. Among these, the motion vectors 24 (for RY) and 25 (for BY) detected by the color difference signal are input to a motion vector comparison circuit 26 which selects the larger one of the motion vectors. The selected motion vector 29 is input to the motion vector selection circuit 28. The motion vector selection circuit 28 receives the motion vector 23 detected by the luminance signal and the output 30 of the image gradient calculation circuit 27 described above. The calculation in the image gradient calculation circuit 27 is obtained according to the following equation.

ΔX = ( _{An + 1, m} − _{An-1, m} ) / 2 ΔY = ( _{An, m + 1} − _{An, m−1} ) / _2Anm indicates an image signal level of n pixels and m lines. . And whether this △ X, 以上 Y is more than the threshold or not (less than)
Is determined. The result 30 is output and the motion vector selection circuit
Give 28.

In the motion vector selection circuit 28, if the input 30 is information equal to or larger than the threshold value, the motion vector selection circuit 28 outputs the motion vector as a motion vector for obtaining a motion vector in a luminance signal. (40) If less than the threshold value, the motion vector 29 detected and selected by the color difference signal is output. (40) This output 40 is simultaneously stored in the motion vector memory 13.

As described above, since the motion vectors of both the luminance signal and the color difference signal are separately detected and selected based on the result of the image gradient calculation, it is possible to detect a motion vector that is extremely faithful to the motion of the image.

FIG. 2 shows a second embodiment of the present invention. Symbol 1 in the figure
12, 13, 14, and 17 are the same as those in FIG. As in the first embodiment, when the six signals 1 to 6 are input, noise and high frequency components are removed by the corresponding two-dimensional LPFs 7 to 12. The signal corresponding to the current field among the outputs is input to the corresponding image gradient calculation circuits 31 to 33.
That is, the difference from the first embodiment is that the signal gradient is first calculated at this point. The calculation is the same as that described in the first embodiment. The calculation result is input to the signal selection circuit 34. It is not necessary to explain that this input is the signal of the current field. The output of the two-dimensional LPFs 7 to 12 is also input to the signal selection circuit 34. Therefore, the signal selection circuit 34 selects a signal based on the input from the image gradient calculation circuits 31 to 33, that is, the calculation result. The selection method selects the luminance signal if the operation result (output of 31) of the luminance signal 1 of the current field is equal to or larger than the threshold, and otherwise selects the larger signal on the color difference signal side (32, 33). However, when the calculation results (32, 33) on the color difference signal side are both equal to or smaller than the threshold value, the luminance signal is selected even if the luminance signal (31) is smaller than the threshold value.

Reference numeral 35 denotes a one-block shift circuit which delays a signal by one block assuming that the operation time of the initial displacement vector selection circuit 14 is one block.

First, a signal is selected as described above, and then input to the initial deviation vector selection circuit 14 and the deviation vector detection circuit 17 and the output is added by the addition circuit 20 to obtain the motion vector 23. It is. The circuit is simpler than in the first embodiment.

Although the first and second embodiments have been described by using a motion vector detection based on a field, it will not be necessary to explain that the present invention can be implemented by replacing the frame with a frame. Although the chrominance signal has been described for RY and BY, it goes without saying that the same can be applied to I, Q, U and V, and the same can be naturally applied to the case of one kind of chrominance signal. In the latter case, there would be no choice among the chrominance signals, so it would be clear that the circuit would be simpler.

(Effects of the Invention) As described above, the present invention employs a method of detecting a motion vector in both a luminance signal and a chrominance signal. Therefore, there is no level change or motion in the luminance signal, and there is no level change or motion in the chrominance signal. In this case, the motion vector can be detected, so that it can be said that a motion vector that is extremely faithful to the motion of the image can be detected. Therefore, an image with very little image distortion can be obtained with a TV system converter using a motion vector.

[Brief description of the drawings]

FIG. 1 is a motion vector detecting circuit according to a first embodiment of the present invention, FIG. 2 is a motion vector detecting circuit according to a second embodiment of the present invention, FIG. 3 is a conventional motion vector detecting circuit, and FIG. FIG. 7 is a diagram for explaining the movement of only the color difference signal. 1 ... luminance signal of current field, 2 ... luminance signal of previous field, 3 ... color difference signal (RY) of current field, 4 ... color difference signal (RY) of previous field, 5 ... current Field color difference signal (BY), 6: Color difference signal (BY) of previous field, 7 to 12: Two-dimensional LPF, 13: Motion vector memory, 14 to 16: Initial displacement vector selection Circuits, 17-19: Gradient operation circuit, 20-22: Addition circuit, 26: Motion vector comparison circuit, 27, 31-33 ... Image gradient operation circuit, 28: Motion vector selection circuit, 34 ... ... signal selection circuit, 35 ... one block shift circuit, 40 ... motion vector signal.

Continued on the front page (72) Inventor Hiroyuki Shimano 1-7-12 Toranomon, Minato-ku, Tokyo Oki Electric Industry Co., Ltd. (72) Inventor Takeo Tsutsui 2-2-1 Jinnan, Shibuya-ku, Tokyo Japan Broadcasting Corporation Inside the Broadcasting Center (72) Yoshihiro Yamamoto 2-2-1 Jinnan, Shibuya-ku, Tokyo Japan Broadcasting Corporation Broadcasting Center (72) Yasushi Inoue ▲ Yoshi ▼ 2-2-1 Jinnan, Shibuya-ku, Tokyo Japan Broadcasting (56) References JP-A-1-309597 (JP, A) JP-A-2-250490 (JP, A) JP-A-58-27489 (JP, A) JP-A-55-162683 (JP, A) JP-A-55-162684 (JP, A) JP-A-60-158786 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name) H04N 11/00-11/24 H04N 9/64-9/78

Claims (2)

(57) [Claims] In a system for detecting a motion vector using two or more fields of a digitalized TV signal, motion vectors are separately detected for a luminance signal and a color difference signal, and an image gradient of the luminance signal is detected. When the calculation result is greater than or equal to a threshold value, the motion vector to be determined is a motion vector detected using a luminance signal, and when the calculation result is less than the threshold value, the motion vector to be determined is a motion vector detected using a chrominance signal. A motion vector detection method for a TV system conversion device, characterized in that: 2. When there are a plurality of types of color difference signals, and when a motion vector to be obtained is a motion vector detected using the color difference signals, a large one of the motion vectors detected using the plurality of types of color difference signals is 2. The motion vector detecting method according to claim 1, wherein the motion vector is obtained.