JPH03284094A - Moving detection signal processing circuit - Google Patents

Abstract

PURPOSE:To detect a fast movement of a picture as movement by providing a spatial filter and magnifying the area of a moving detection signal only when a difference signal between a picture element and its surrounding movement detection signal is a prescribed value or below. CONSTITUTION:A 1H delay circuit 4, a 1st selection circuit 10, an adder 12, a 2nd selection circuit 14 and a discrimination circuit 16 form a spatial filter means magnifying an area of a movement detection signal inputted to a terminal 2 in response to an output of an absolute value circuit 8. Through the constitution above, when a moving detection signal generated resulting from a noise or a steep edge of a picture detected as the movement is fed to the input terminal 2, since an output of the absolute value circuit 8 is larger than a prescribed value, the area of the moving detection signal is not magnified, but only when the difference signal is a prescribed value or below, the area of the moving detection signal is magnified. Thus, pulse noise or picture steep edge is not subjected to magnification processing in mistake as the movement and the fast movement of the picture is detected as the movement.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image processing apparatus such as a television receiver, and particularly to a motion detection signal processing circuit that detects motion of an image.

[Prior Art] Currently, the NTSC system used in television broadcasting etc. is a composite video signal in which the 71111 degree signal component and the color signal component are frequency multiplexed. If the luminance signal and color (No. 8) are not separated, crosstalk and resolution degradation will occur.
The method uses 2:1 interlaced scanning, which causes problems such as line flicker and a decrease in vertical resolution.

Therefore, in order to accurately separate the luminance signal and color (ff signal), interpolate the lines thinned out in the inklay input signal, and convert them into sequential scanning (non-interlaced) equal signals.
Motion adaptive processing is being considered, which changes the processing method depending on moving parts and stationary parts of an image.

In motion adaptive processing, it is important to detect the moving part of the image. In particular, if moving parts cannot be detected and still processing is performed, a significant deterioration in image quality will occur. Therefore, in the motion detection circuit, even if a stationary part may be mistakenly detected as moving, it is necessary to prevent a moving part from being mistakenly detected as stationary.

JP-A No. 63-90987 proposes a circuit that prevents erroneously detecting a moving part as stationary. In addition to detecting the movement of the low frequency component of the luminance signal, the movement of at least the color signal band is detected based on the difference signal between two frames, and the movement detection signal is processed by a spatio-temporal filter circuit to detect the motion area. The image is enlarged and smoothed to ensure that no movement is overlooked.

[Problems to be Solved by the Invention'] The expansion of the motion region using the spatio-temporal filter circuit disclosed in Japanese Patent Laid-Open No. 63-90987 has a first-order problem.

First, if the S/N of the signal is poor, a motion point that occurs as an isolated point due to pulse noise will be erroneously detected as motion because its area will be enlarged by spatiotemporal processing.

Furthermore, if one image moves slightly due to camera vibration or telecine, the difference signal will be significant at steep edge parts of the image, even if the movement of the edge part is sufficiently small relative to the sample interval. Therefore, by expanding the region, even parts that can be processed still will be mistakenly detected as motion.

Conversely, if the image moves quickly, the curse field is not filtered in the vertical direction, so in the first field where movement occurs, the movement area is insufficiently enlarged and cannot be detected.

In particular, the above two problems of false detection become more noticeable when the motion detection sensitivity is increased in order to prevent motions that should be detected from not being detected, resulting in poor motion detection accuracy.

The present invention provides a motion detection signal processing circuit that is capable of detecting fast motion of an image as motion without erroneously detecting pulse noise or small movement portions of steep edges of the image as motion. The purpose is to

[Means for Solving the Problems] A motion detection signal processing circuit of the present invention is a motion detection signal processing circuit that processes motion detection No. 1m based on interframe differences of image signals. Calculating means for calculating a difference signal between the motion detection signal and the motion detection signals of pixels surrounding the pixel to be subjected to motion adaptive processing, and expanding the area of the motion detection signal of the pixel to be subjected to motion adaptive processing according to the difference signal. and spatial filter means.

[Operation] In the motion detection according to the present invention having the above configuration (in the i-processing circuit, the motion detection signal of a pixel detected as motion due to pulse noise, an image, or a steep edge part is detected as a motion detection signal of one surrounding pixel). Since the difference signal (No. 1) of the calculation means is larger than a predetermined value, the spatial filter means does not expand the area of the motion detection signal.

Therefore, pulse noise and steep edge portions of the image will not be mistakenly detected as motion.

When the difference signal is less than a predetermined value, the spatial filter means expands the area of the motion detection signal, so that fast motion of the image is detected as motion.

[Embodiment] FIG. 1 shows the configuration of an embodiment of the motion detection (d-processing circuit) of the present invention.The input terminal 2 receives a composite video signal in which a luminance signal component and a color signal component are frequency-multiplexed. 1 frame f
A luminance signal low frequency component motion detection signal derived based on the low frequency component of the ill difference signal, and a chrominance signal band motion detection f-scale derived based on the difference signal between two frames of the composite video signal. A motion detection signal obtained by combining the two is supplied. LH
The delay circuit 4 delays the input motion detection (No. 1 by one horizontal scanning period) and outputs it.The subtracter 6 (calculation, means)
A motion detection signal input to terminal 2 (this corresponds to a motion detection signal of a pixel that will undergo motion adaptive processing from now on),
11-2 Outputs a difference signal from the output signal of the delay circuit 4 (this corresponds to the motion detection signal of the surrounding pixels of the pixel to be subjected to motion adaptive processing). The absolute value circuit 8 outputs the absolute value of this difference signal.

The first selection circuit 10 compares the motion detection signal input to the terminal 2 and the output signal of the IH delay circuit 4, and outputs the signal No. 4m having the smaller value. The adder 12 adds the motion detection signal input to the terminal 2 and the output signal of the IH delay circuit 4, multiplies the result by 1/2, and outputs the result. The second selection circuit 14 outputs either the output of the first selection circuit 10 or the output of the adder 12 to the terminal 18 as a motion detection signal based on the output of the determination circuit 16.

When the output of the absolute value circuit 8 is larger than a predetermined value, the determination circuit 16 instructs the second selection circuit 14 to select and output the output of the first selection circuit 10, so that the output of the absolute value circuit 8 is larger than the predetermined value. When it is less than the value, the second selection W114 is instructed to select and output the output of the adder 12.

The LH delay circuit 4, the first selection circuit 10, the adder 12, the second J selection circuit 14, and the determination circuit 16 expand the area of the motion detection signal input to the terminal 2 according to the output of the absolute value circuit 8. Construct a spatial filter means to carry out. The III delay circuit 4 and the adder 12 directly constitute a vertical filter that expands the area of the motion detection signal. The output terminal 18 of the second selection circuit 14 is
A spatio-temporal filter circuit connected to a spatio-temporal filter circuit (not shown) as disclosed in No. 0987 expands the area of the motion detection signal and provides it to a motion adaptive processing circuit (not shown).

In the embodiment of the ft51 diagram configured as described above, a motion detection signal of a pixel to be subjected to motion adaptive processing, which is generated as a result of detecting pulse noise or a steep edge portion of an image as motion, is input to the input terminal. 2, the value of this signal is significantly larger than the value of the pixel motion detection signal output from the IH delay circuit 4 one horizontal scanning period ago.
The first selection circuit 10 selects the output of the IH delay circuit 4.

Also.

Since the difference signal output by the subtracter 6 and therefore the absolute value output by the absolute value circuit 8 becomes larger than the predetermined value, the determination circuit 16
instructs the second selection circuit 14 to select and output the output of the first selection circuit 10.

Therefore, the area of the motion detection signal supplied to input terminal 2 is not expanded (that is, the motion detection signal is not filtered in the vertical direction), and pulse noise and steep edges of the image are mistakenly interpreted as motion. There will be no enlargement processing.

If the difference between the motion detection signal of the pixel to be subjected to motion adaptive processing supplied to the input terminal 2 and the motion detection signal of the pixel one horizontal scanning period before output from the IH delay circuit 4 is less than a predetermined value, the determination is made. The circuit 16 instructs the second selection circuit 14 to select and output the output of the adder 12. Therefore, the region of the motion detection signal supplied to the input terminal 2 is expanded (that is, the motion detection signal is filtered in the vertical direction), so that even fast motion in one image can be detected as motion.

A motion detection signal corresponding to pulse noise or a steep edge portion of an image is output from the delay circuit 4 after one horizontal scanning period after being input to the terminal 2. At this time, the first selection circuit 10 selects the signal input from input terminal 2. Furthermore, since the output of the absolute value circuit 8 is large, the second selection circuit 14 selects the output of the first selection circuit 10.

FIG. 2 shows the configuration of another embodiment of the motion detection signal processing circuit of the present invention. The input terminal 102 is supplied with the same motion detection in signal as in the embodiment of FIG.

The III delay circuit 104 converts the input motion detection signal into 1
Output is delayed by the horizontal scanning period. The subtracter 106 extracts the motion detection signal input to the terminal 102 (this corresponds to the motion detection signal of the pixel that will undergo motion adaptive processing) and the output signal of the 11-1 delay circuit 104 (this corresponds to the motion detection signal of the pixel that will be motion adaptively processed). (corresponding to motion detection signals of surrounding pixels) of the pixel to be adaptively processed.

The determination circuit 11G assigns 1 to the coefficient that changes complementary to the difference signal output from the subtracter 10'6 (calculating means).
and a signal indicating 2 is generated by the coefficient multiplier 120, respectively.
and output to 122. The coefficient unit 120 is connected to the terminal 102
The coefficient of the motion detection signal input to the input signal is multiplied by 1. The coefficient unit 122 1) multiplies the output of the I delay circuit 104 by a coefficient of 2; The coefficients 1 and 2 satisfy the relationship 2=1 for K1+, and when the difference signal output from the subtracter 106 is 0, 1=2=0.5, and the difference signal is positive and the absolute value It is determined that 1 approaches 0 as the difference signal becomes larger, and 1 approaches 1 as the difference signal becomes negative and the absolute value becomes larger. Namely.

Motion detection signal input to terminal 102 and IH delay circuit 1
04, and if the difference is large, it is determined to be noise, and the motion detection signal input to the terminal 102 and the I
The coefficient of the coefficient multiplier corresponding to the smaller signal among the motion detection signals output from the H delay circuit 104 is increased.

FIG. 3 shows an example of characteristics for a difference signal with a coefficient of 1 output by the determination circuit 116. The adder 124 is a coefficient unit 1
The outputs of 20 and 122 are added together and outputted to terminal 118 as a motion detection signal.

IH delay circuit 104, determination circuit 116, coefficient unit 120
and 122, and adder 124 is subtracter 106
A spatial filter means is configured to expand the area of the motion detection signal input to the terminal 102 in accordance with the difference signal output from the terminal 102. The output terminal 11B of the adder 124 is
A spatio-temporal filter circuit connected to a spatio-temporal filter circuit (not shown) as disclosed in Japanese Patent Application No. 3-90987 expands the area of the motion detection signal and converts it into a motion adaptive processing circuit (
(not shown).

In the embodiment of FIG. 2 configured as described above, when a motion detection fi signal corresponding to pulse noise or a steep edge portion of an image is supplied to the input terminal 102, this signal value is The difference signal output from the lever subtracter 106, which is significantly larger than the value of the motion detection signal of the pixel one horizontal scanning period before output from the delay circuit 104, is positive and has a large absolute value.

As a result, the coefficient output from the determination circuit 116 approaches 0, and the coefficient multiplier 120 multiplies the motion detection signal input to the terminal 102 by O or a value close to 0.

Therefore, the region of the motion detection signal supplied to the terminal 102 is not enlarged, and pulse noise or steep edge portions of the image are not erroneously enlarged as motion.

Furthermore, a motion detection signal corresponding to pulse noise or a steep edge portion of an image is output from the IH delay circuit 104 one horizontal scanning period after being input to the terminal 102.
At this time, the difference signal output from the subtracter 106 is negative and its absolute value becomes large, so the coefficient 2 approaches 0 in the coefficient output from the decision port Wt 10 G, and the coefficient unit 122 outputs it from the 1H delay circuit 104. Multiply the signal by 0 or a value close to O.

Therefore, the region of the motion detection signal output from the LH delay circuit 104 is not expanded, and pulse noise and movement in steep edge portions of the image are not erroneously expanded or processed.

If the difference between the motion detection signal of the pixel to be subjected to motion adaptive processing supplied to the terminal 102 and the motion detection signal of the pixel one horizontal scanning period before outputted from the IH delay circuit 104 is 0 or small, the judgment circuit 116 The coefficients 1 and 2 are set to the same value or a close value. Therefore, the motion detection signal supplied to the terminal 102 and the output signal of the LH delay circuit 104 are multiplied by the same value or a similar value, and are added by the adder 124, so that the area of the motion detection signal is expanded. Fast movements in the image are detected as motion.

[Effects of the Invention] As described above, the motion detection signal processing circuit of the present invention prevents pulse noise or steep edge portions of an image from being erroneously enlarged as motion, and also prevents fast movement of images. can be detected as movement.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of one embodiment of the motion detection signal processing circuit of the present invention, and FIG. 2 is a block diagram showing the structure of another embodiment of the motion detection signal processing circuit of the present invention. 3
The figure is a graph showing an example of the relationship between the coefficient 1 output by the determination circuit of the embodiment shown in FIG. 2 and the difference signal. 4.104... IH delay circuit, 6,106... Subtractor (calculating means), 8... Absolute value circuit, 10... First selection circuit, 12,124... Adder, 14... Second selection circuit, 16.116... Judgment circuit, 120.1
22...Coefficient unit.

Claims (1)

[Claims] In a motion detection signal processing circuit that processes a motion detection signal based on an inter-frame difference of an image signal, a motion detection signal of a pixel of the image to be subjected to motion adaptive processing and motion detection of pixels surrounding the pixel to be subjected to motion adaptive processing. 1. A motion detection signal processing circuit comprising: arithmetic means for calculating a difference signal between the signals; and a spatial filter means for expanding the area of a motion detection signal of a pixel subjected to motion adaptive processing according to the difference signal.