JP3407429B2 - Video signal processing device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a video signal processing device such as a flicker correction circuit, a camera shake correction circuit and a noise reduction circuit, which is used in video equipment such as a video camera for the purpose of improving image quality.

[0002]

2. Description of the Related Art Recently, in video cameras, there is an increasing demand for high-quality recording by a simple operation, and various techniques have been introduced for this. Among these are flicker correction, camera shake correction, and S /
N improvement etc. are mentioned, but these are explained.

First, flicker correction will be described. Flicker is a wide range of flicker on the screen,
Here, the thing that the screen brightness changes periodically, which appears when an image is taken under a fluorescent lamp in the East Japan region, will be described. Since the commercial power source is 50 Hz in the eastern Japan area, the fluorescent lamp blinks at 100 Hz. But NT
Since the sampling frequency in the time direction in the SC type video camera is 60 Hz, the screen brightness changes in a cycle of 20 Hz due to this difference. This will be described with reference to FIG. FIG. 9 shows a state in the frequency domain when a 100 Hz signal is sampled at 60 Hz. Nth harmonic of the sampled signal (n = 1,2,3, ...)
Appears at a position 100 Hz above and below a frequency n times the sampling frequency 60 Hz. Focusing on the region of 30 Hz or less, which is the band in the time direction, it is found that flicker occurs when 20 Hz of the second harmonic is mixed in the band.

As a method of removing this flicker, CC
There is a method of setting the exposure time of D to 1/100 second and synchronizing it with the blinking of the fluorescent lamp.
The sensitivity is reduced as compared with the case where the exposure time is 0 second.

Therefore, there is a method of changing the gain for the input signal according to the phase of the 20 Hz component extracted from the input video signal. This will be described.
FIG. 10 (1) shows a state in which a flicker of 20 Hz is mixed in a signal of a constant level. Since the output signal of the CCD is 60 Hz, one cycle of 20 Hz corresponds to three cycles of 60 Hz, and as shown in the figure, the signal has three levels of a, b, and c. If this signal is multiplied by a coefficient as shown in (2) of the figure, the signal can be maintained at a substantially constant level as shown in (3) of the figure, and the flicker can be removed.

Next, camera shake correction will be described. "Shake" refers to a state in which a small shake of the hand is transmitted to the camera when the user holds the video camera by hand, and the entire screen vibrates in small steps. Currently, there are roughly two types of methods for correcting this, an optical type and an electronic type. Here, the electronic type will be briefly described. FIG. 11 shows the principle of electronic image stabilization. When the amount and direction (motion vector) of the motion of the subject is detected from the image sensor output signal, and at the same time the image sensor output signal is stored in the memory and it is determined that "camera shake", the read address of the memory is changed according to the motion vector. The signal stored in the memory is “cut out” by moving the start point and the end point and reading the data. By the above processing, the movement of the entire screen due to camera shake is absorbed, and the subject becomes almost stationary.

Here, a method is often used in which the memory for storing the image sensor output signal is used as a field memory, and the motion vector of the object is detected over one field, and then the content of the memory is read out in the next field. As another method, a line memory is used instead of the field memory, and a motion vector is obtained by prediction from past results to reduce cost.

Finally, S / N improvement, that is, noise reduction processing will be described. There are various methods for reducing noise at present, but the three-dimensional noise reduction process using a field or frame memory is the most effective.

Explaining this processing, the video signal has a high correlation in the time direction, and conversely the noise has a small correlation. Therefore, by adding and averaging a plurality of consecutive fields or frames, the noise level can be reduced with almost no change in the video signal.

However, since the field memory and the frame memory are expensive at present, an IIR (Infinite Impulse Response) filter is generally used as the addition processing in the time direction. Figure 1 shows the specific noise reduction circuit configuration.
2 shows. Reference numeral 9 indicates a multiplication means of k (0 ≦ k ≦ 1). The processing of this circuit is expressed by an equation (Equation 1).

[0011]

[Equation 1]

Where x _n is the input signal, y _n and y _n-1
Are current and 1 field (or frame) respectively
The previous output signal is shown. This processing is called field (or frame) cyclic noise reduction processing. The field cyclic processing will be described below as an example. By transforming (Equation 1), (Equation 2) is obtained.

[0013]

[Equation 2]

A circuit configuration based on this equation is shown in FIG. From the figure, it can be seen that in the three-dimensional noise reduction processing, it can be considered that the field difference is multiplied by the coefficient k and subtracted from the input signal.

Next, consider the coefficient k. In a still image, the contents of the field difference are all noise,
This can be removed by subtracting noise from the input signal with k = 1. However, in the moving image, the contents of the field difference are noise and a moving part of the signal, and when k = 1 is subtracted from the input signal, image quality deterioration called an afterimage is caused. Therefore, it is necessary to distinguish noise from the moving part of the signal by some method.

What is often used to distinguish between noise and a signal is that a signal generally has a higher level than noise, and the level of the field difference input according to the characteristics shown in FIG. 14 is used. Is small, the output value is judged to be a “still image part” and the output value is equal to the input value (k =
1), the output level is made smaller than the input level as the input level becomes higher as shown in (a) of the figure, so that k <1, and when the input level is higher than a predetermined level T, it is determined to be a “moving image part”. Output is set to 0 (k = 0), or a constant value is output when the input level is higher than a predetermined level T as shown in FIG. Make k asymptotic to 0. This processing is called non-linear processing because the input / output characteristics are non-linear.

In addition to the above, as a method for distinguishing noise from a signal, there is a method that generally utilizes the fact that a signal has a lower frequency than noise. As shown in FIG. 15, the field difference is divided into a plurality of bands, and the above-described non-linear processing is independently performed for each band. 9a and 9b are shown in FIG.
4 means a non-linear process having the characteristics shown in FIG.
It is the same processing as the multiplication means in 2 and 13. In the circuit having such a configuration, the nonlinear processing means 9
The level T at a and 9b has a small low frequency component containing a large amount of signals and a large high frequency component containing a large amount of noise.

[0018]

However, when the input to the field recursive noise reduction processing circuit as described above is a signal accompanied by flicker as described above, the "still image section" is used.
Is erroneously determined to be a “moving image portion” and becomes a value of k = 0 or close to 0, and the noise reduction effect is reduced. This will be described.

Consider a case in which a signal accompanied by flicker is input to a noise reduction circuit having a non-linear processing means having the characteristic shown in FIG. A signal accompanied by flicker is a signal that should originally have a constant level as shown in FIG. 10, but changes with time. The width of this level fluctuation is directly obtained by the noise reduction circuit as a difference from the signal one field before. If the width of this level fluctuation is smaller than the level T shown in FIG. 14A in the non-linear processing means, noise reduction processing is performed, but if it is larger than T, noise reduction is not performed at all. .

Further, when the signal accompanied by the above-mentioned camera shake is inputted, the moving image is inputted to the noise reduction circuit as a moving image of the whole screen although the subject is stationary. 14 in the non-linear processing means
The noise reduction effect becomes smaller as the level T shown in FIG.

The present invention solves the above problems, and provides a video signal processing device which can obtain a sufficient noise reduction effect even when a signal accompanied by flicker or camera shake is input.

[0022]

In order to achieve this object, a video signal processing apparatus according to the first aspect of the present invention comprises a filter means for extracting a predetermined frequency from an input signal and a level variation of the output of the filter means. From the level fluctuation correction unit output, a level fluctuation correction unit including a gain calculation unit that calculates a gain and a multiplication unit that calculates a product of an input signal and the gain calculation unit, a storage unit that stores a signal for a predetermined period, and a level fluctuation correction unit output. First subtraction means for subtracting the output of the storage means;
A non-linear means for performing non-linear processing on the output of the subtracting means, and a second subtracter for subtracting the non-linear means output from the level fluctuation correction section output and using the output as the input of the storage means.
The noise reduction unit is composed of

In order to achieve the above-mentioned object, the video signal processing device of the second invention is a motion vector detecting means for detecting and outputting the direction and amount of the motion of the signal from the input video signal, and the input video signal. First storage means for storing for a predetermined period of time, and control means for outputting a part of the signal stored in the first storage means based on the output of the motion vector detection means,
An image stabilization unit, a second storage unit that stores a signal for a predetermined period, a first subtraction unit that subtracts an output of the second storage unit from an output of the image stabilization unit, and a first subtraction unit output. Configuration of a noise reduction unit composed of a non-linear means for performing non-linear processing on the other hand, and a second subtractor which subtracts the non-linear means output from the shake correction section output and uses the output as the input of the second storage means. have.

In order to achieve the above object, the video signal processing device of the third invention is a motion vector for predicting and detecting the direction and amount of motion of a signal from an input video signal and outputting the predicted value and the detected value. Prediction detecting means, image sensor driving means for controlling the driving of the image sensor based on the output of the motion vector prediction detecting means, first storing means for storing the input video signal for a predetermined period, and prediction output of the motion vector predicting detecting means. A first control means for outputting a part of the signal stored in the first storage means based on
A shake correction section, a second storage means for storing a signal for a predetermined period, a first subtraction means for subtracting the output of the second storage means from the output of the first shake correction section, and a first subtraction means output. Noise reduction, which comprises a non-linear means for performing non-linear processing, and a second subtracter which subtracts the non-linear means output from the first camera shake correction section output and uses the output as the input of the second storage means. And a second storage means, and a second control for outputting a part of the signal stored in the second storage means based on the detection output of the motion vector prediction detection means.
And a second camera shake correction section composed of the control means of FIG.

[0025]

According to the first aspect of the present invention, with the above-mentioned configuration, the flicker correction is performed on the input signal accompanied by the flicker to improve the noise reduction effect.

Further, the second and third aspects of the present invention have the above-mentioned structure to perform camera shake correction on an input signal accompanied by camera shake to improve the noise reduction effect.

[0027]

Embodiments of the video signal processing device of the present invention will be described below.
A description will be given with reference to the drawings.

FIG. 1 is a block diagram of a video signal processing device according to a first embodiment of the first invention. First, the input video signal input from the input terminal 20 is BPF (Ba
nd Pass Filter) 2, gain calculator 3, and multiplier 4
Is input to the flicker correction unit 1. The BPF 2 extracts the 20 Hz component included in the input video signal. In the 20 Hz component obtained here, as shown in FIG. 10A, data having three types of levels a, b, and c are arranged at 1 field (1/60 second) intervals.

The gain calculator 3 obtains the arithmetic mean value of the three levels of the 20 Hz component and divides it by the respective data of a, b, and c to determine the correction gain for each data. That is, the gain is determined for each of a, b, and c as shown in (Equation 3).

[0030]

[Equation 3]

The multiplier 4 multiplies the input video signal by the gain output from the gain calculator 3. As a result, the corrected level becomes (Equation 4), which can be a constant value.

[0032]

[Equation 4]

The output of the flicker correction unit 1 is subsequently input to the noise reduction unit 5 including the first and second subtractors 6 and 7, the field memory 8 and the non-linear processing means 9. The input signal to the noise reduction unit 5 is input to the first and second subtractors 6 and 7. In the first subtractor 6, the difference between the input signal and the output of the field memory 8 is obtained and input to the non-linear processing means 9. The non-linear processing means 9 has an input / output characteristic as shown in FIG.
When the input level is low, the input is output as it is to set k = 1, and as the input level increases, the output level becomes lower than the input level to set k <1, and when the input level is higher than the predetermined level T, Output 0 and set k = 0. Assuming that the flicker is sufficiently removed in the flicker correction unit in the previous stage, the output level of the first subtractor 6 in the stationary portion of the screen is as shown in FIG.
Since it is considered that the level is below the level T shown in (a), the output of the non-linear processing means 9 does not become 0, and an appropriate value is output.

The output of the non-linear processing means 9 is the second subtractor 7
At, the signal is subtracted from the input signal to obtain a noise reduced output. The output of the second subtractor 7 is input to the field memory 8.

Next, FIG. 2 shows a block diagram of a video signal processing apparatus according to a second embodiment of the first invention. Input terminal 2
The input video signal input from 0 is a flicker correction unit 1 including a BPF 2, a gain calculator 3, and a multiplier 4.
Entered in. The operation of the flicker correction unit 1 is exactly the same as that of the first embodiment, and therefore its explanation is omitted.

The output of the flicker correction unit 1 is subsequently supplied to the first and second subtractors 6 and 7, the field memory 8 and the LP.
F (Low Pass Filter) 10, HPF (High Pass Filte)
r) 11, non-linear processing means 9a and 9b, combining means 1
It is input to the noise reduction unit 5 composed of two. The input signal to the noise reduction unit 5 is the first and second subtractors 6 and 7
Entered in. In the first subtractor 6, the difference between the input signal and the output of the field memory 8 is obtained. The output of the first subtractor 6 is input to the LPF 10 and the HPF 11. The outputs of the LPF 10 and HPF 11 are input to the non-linear processing means 9a and 9b. Non-linear processing means 9
As a and 9b, those having an input / output characteristic as shown in FIG. 14A are used. Here, if it is assumed that the flicker is sufficiently removed by the flicker correction unit in the previous stage, it is considered that the output levels of the LPF 10 and the HPF 11 are below the level T shown in FIG. , Non-linear processing means 9a and 9
The output of b does not become 0, but an appropriate value is output.

As described above, since the LPF 10 output mainly contains the video signal and the HPF 11 output mainly contains the noise, the predetermined level T in the non-linear processing means 9a is higher than that in the non-linear processing means 9b. By setting it small, it is possible to enhance the noise reduction effect while suppressing the afterimage.

The outputs of the non-linear processing means 9a and 9b are combined by the combining means 12 into two frequency components,
Is subtracted from the input signal to obtain an output with reduced noise. The output of the second subtractor 7 is input to the field memory 8.

Next, a video signal processing apparatus according to the first embodiment of the second invention will be described with reference to the drawings. FIG. 3 is a block diagram of a video signal processing device according to the first embodiment of the second invention. The input video signal input from the input terminal 20 is the motion vector detector 1
4, first field memory 15, memory control means 16
Is input to the image stabilization unit 13.

The input video signal is input to the motion vector detector 14 and written in the first field memory 15. The motion vector detector 14 detects and outputs a motion vector in the input video signal from changes in data at a plurality of representative points provided on the screen. The output of the motion vector detector 14 is input to the memory control means 16.
The memory control unit 16 changes the read control signal of the first field memory 15 in accordance with the motion vector, and moves the read address area. For example, when the direction of the motion vector is rightward, the horizontal read address start point and end point of the first field memory 15 are incremented, and when the direction of the motion vector is upward, the vertical read address is read. Decrement start and end points. By the above processing, the camera shake is removed from the input video signal, and a signal in a substantially stationary state is obtained.

The output of the camera shake correction unit 13 is subsequently supplied to the first and second subtractors 6 and 7, and the second field memory 2.
3 and is input to the noise reduction unit 5 including the non-linear processing means 9. The input signal to the noise reduction unit 5 is the first and second signals.
Is input to the subtracters 6 and 7. In the first subtractor 6, the difference between the input signal and the output of the second field memory 23 is obtained and input to the non-linear processing means 9. The non-linear processing means 9 has, for example, an input / output characteristic as shown in FIG. 14A. When the input level is small, the input is output as it is to set k = 1, and the output level increases as the input level increases. When the input level is smaller than the input level, k <1 is set. When the input level is larger than the predetermined level T, 0 is output and k = 0. Here, assuming that the camera shake correction unit in the preceding stage has sufficiently removed the camera shake, it is considered that the output level of the first subtractor is within the level T shown in FIG. Therefore, the output of the non-linear processing means 9 does not become 0, and an appropriate value is output.

The output of the non-linear processing means 9 is the second subtractor 7
At, the signal is subtracted from the input signal to obtain a noise reduced output. The output of the second subtractor 7 is input to the second field memory 23.

Next, FIG. 4 shows a block diagram of a video signal processing apparatus according to a second embodiment of the second invention. Input terminal 2
The input video signal input from 0 is the motion vector detector 14, the first field memory 15, the memory control means 1
It is input to the camera shake correction unit 13 composed of 6. The operation of the camera shake correction unit 13 is the same as that in the first embodiment.
The description is omitted because it is exactly the same as.

The output of the camera shake correction unit 13 is continuously output to the first and second subtractors 6 and 7, and the second field memory 2.
3, LPF 10, HPF 11, non-linear processing means 9 a and 9 b, and synthesizing means 12 are input to the noise reduction unit 5. The input signal to the noise reduction unit 5 is the first and second signals.
Is input to the subtracters 6 and 7. In the first subtractor 6, the difference between the input signal and the output of the second field memory 23 is obtained. The output of the first subtractor 6 is input to the LPF 10 and the HPF 11. LPF10, HPF11
The respective outputs are input to the non-linear processing means 9a and 9b. The non-linear processing means 9a and 9b are shown in FIG.
A device having an input / output characteristic as shown in (a) is used. Here, assuming that the camera shake correction unit in the preceding stage has sufficiently removed the camera shake, LPF1
Since it is considered that the output levels of 0 and HPF 11 are below the level T shown in FIG. 14A, the outputs of the non-linear processing means 9a and 9b do not become 0, and appropriate values are output. With such a configuration, the field difference is divided into a plurality of bands and processed independently as described above, and it is possible to enhance the noise reduction effect while suppressing the afterimage.

As described above, there are two methods for obtaining the motion vector in the camera shake correction unit 13. One is to detect a motion vector from a video signal for one field and output it after one field. The other is 1
The motion vector for the current field is predicted from the motion vector detection result before the field, and when the input of the current field is completed, the motion vector is determined and used for prediction for the next field. Here, if the former is called a deterministic method and the latter is called a predictive method, a field memory is required as a memory means in the deterministic method. However, in the case of the prediction method, for example, a CCD having a large number of pixels is used, and at the time of outputting a signal from this CCD, a scanning line signal outside the necessary range based on the motion vector obtained by prediction is discarded by high-speed sweeping. If "cutting" is performed in the vertical direction, then only "cutting" is performed in the horizontal direction, and a line memory may be used as the memory means.
Although this method is somewhat difficult in terms of correction accuracy, it has the advantage that the circuit scale can be reduced. This is shown below.

FIG. 5 shows a block diagram of a video signal processing apparatus according to a third embodiment of the second invention. To explain the operation, the input video signal input from the input terminal 20 is input to the image stabilization unit 13 including the motion vector predictor 25, the CCD drive pulse generator 17, the line memory 22, and the memory control unit 16.

The input video signal is input to the motion vector predictor 25 and written in the line memory 22. The motion vector predictor 25 predicts and outputs a motion vector for the current field from the result of motion vector detection one field before, and the output is input to the CCD drive pulse generator 17 and the memory control means 16.

The CCD drive pulse generator 17 changes the CCD drive signal according to the vertical component of the motion vector,
The scanning line signal of an unnecessary portion is discarded by high-speed sweeping. For example, in the case of the NTSC system, 2 per field
2 from PAL CCD that can obtain 88 scanning line signals
Considering to cut out 40 lines, normally, 22 scanning line signals are first discarded by high-speed sweeping, 240 signals are read out, and the remaining 22 lines are discarded. Next, when the direction of the motion vector is upward, the number of scanning lines that are swept out at high speed before reading is reduced to virtually decrement the vertical read address start point and end point on the CCD.

The memory control means 16 changes the read control signal of the line memory 15 according to the horizontal component of the motion vector and moves the read address area. For example, when the direction of the motion vector is rightward, the read address start point and end point of the line memory are incremented. By the above processing, the camera shake is removed from the input video signal, and a signal in a substantially stationary state is obtained.

The output of the camera shake correction unit 13 is subsequently input to the noise reduction unit 5 including the first and second subtractors 6 and 7, the field memory 23, and the non-linear processing means 9. Since the operation is exactly the same as that of the first embodiment, its explanation is omitted.

Next, FIG. 6 shows a block diagram of a video signal processing apparatus according to a fourth embodiment of the second invention. Input terminal 2
The input video signal input from 0 is the motion vector detector 14, the CCD drive pulse generator 17, the line memory 1
5, input to the camera shake correction unit including the memory control unit 16. The operation of the camera shake correction unit is exactly the same as that of the camera shake correction unit in the third embodiment, and therefore the description thereof is omitted.

The output of the camera shake correction unit 13 is subsequently supplied to the first and second subtractors 6 and 7, the field memory 8 and the LP.
F10, HPF11, non-linear processing means 9a and 9b,
It is input to the noise reduction unit 5 including the synthesizing unit 12. Although the operation of the noise reduction unit 5 is exactly the same as that of the second embodiment, the description thereof will be omitted. However, by dividing the field difference into a plurality of bands and processing them independently, the noise reduction effect can be enhanced while suppressing the afterimage. It will be possible.

As described above, if the camera-shake correction unit is the motion vector determination method, there are problems in terms of circuit scale, and if it is the prediction method, there are problems in correction accuracy. Therefore, the following is a configuration configured to compensate for each of the disadvantages.

FIG. 7 is a block diagram of a video signal processing device according to the first embodiment of the third invention. The input video signal input from the input terminal 20 includes a motion vector prediction detector 26, a CCD drive pulse generator 17, a line memory 2
2, input to the first camera shake correction unit 13 including the first memory control unit 24. The operation of the first camera shake correction unit 13 is exactly the same as that of the camera shake correction unit 13 in the second aspect of the present invention, and thus the description thereof is omitted.

The output of the first camera shake correction unit 13 is subsequently supplied to the noise reduction unit 5 including the first and second subtractors 6 and 7, the 3-port field memory 23, and the non-linear processing means 9.
Entered in. Here, the 3-port field memory has one write port and two read ports, and each port can read and write asynchronously. The input signal to the noise reduction unit 5 is input to the first and second subtractors 6 and 7. In the first subtractor 6, the difference between the input signal and the output from the first read port of the 3-port field memory 23 is obtained and input to the non-linear processing means 9. The non-linear processing means 9 is shown in FIG.
It has the input / output characteristics as shown in (a), and when the input level is small, the input is output as it is so that k = 1, and as the input level increases, the output level becomes smaller than the input level. If <1 and the input level is higher than the predetermined level T, 0 is output and k = 0. Here, assuming that the camera shake correction unit in the preceding stage has sufficiently removed the camera shake, it is considered that the output level of the first subtractor is within the level T shown in FIG. Therefore, the output of the non-linear processing means 9 does not become 0, and an appropriate value is output.

The output of the non-linear processing means 9 is the second subtractor 7
At, the signal is subtracted from the input signal to obtain a noise reduced output. The output of the second subtractor 7 is input to the 3-port field memory 23.

The output signal of the noise reduction section 5 is input to the second camera shake correction section 18 including the 3-port field memory 23 and the second memory control means 19. The determined motion vector output from the motion vector prediction detector 26 included in the first camera shake correction unit 13 is input to the second memory control unit 19. Second memory control means 19
Changes the second read port control signal of the 3-port field memory 23 according to the determined motion vector, and moves the read address area. For example, when the direction of the motion vector is rightward, the horizontal read address start point and end point of the 3-port field memory 23 are incremented, and when the direction of the motion vector is upward, the vertical read address start is started. Decrement the point and end point. By the above processing, the camera shake is removed from the input video signal, and a signal in a substantially stationary state is obtained. With such a configuration, the field memories can be integrated into one, and since the final “cutout” uses the determined motion vector, the correction accuracy is sufficient.

Finally, FIG. 8 shows a block diagram of a video signal processing apparatus according to the second embodiment of the third invention. The input video signal input from the input terminal 20 is input to the first image stabilization unit 13 including the motion vector prediction detector 26, the CCD drive pulse generator 17, the line memory 22, and the first memory control means 16. . The operation of the first camera shake correction unit 13 is exactly the same as that of the first embodiment, and therefore its explanation is omitted.

The output of the first camera shake correction unit 13 is subsequently a noise composed of the first and second subtractors 6 and 7, the 3-port field memory 23, the LPF 10, the HPF 11, the non-linear processing means 9a and 9b, and the synthesizing means 12. Reduction unit 5
Entered in. The input signal to the noise reduction unit 5 is input to the first and second subtractors 6 and 7. In the first subtractor 6, the input signal and the 3-port field memory 2
The difference of the outputs of the first read ports of 3 is determined. The output of the first subtractor 6 is input to the LPF 10 and the HPF 11. The outputs of the LPF 10 and HPF 11 are input to the non-linear processing means 9a and 9b. As the non-linear processing means 9a and 9b, for example, those having an input / output characteristic as shown in FIG. 14 (a) are used. Here, assuming that the camera shake correction section in the preceding stage has sufficiently removed the camera shake, the LPF 10 and the HPF 11 are displayed in the still section of the screen.
Since it is considered that the output level of the above is less than the level T shown in FIG. 14A, the outputs of the non-linear processing means 9a and 9b do not become 0, and an appropriate value is output. With such a configuration, the field difference is divided into a plurality of bands and processed independently, and the noise reduction effect can be enhanced while suppressing the afterimage.

The output signal of the noise reduction unit is input to the second camera shake correction unit 18 including the 3-port field memory 23 and the second memory control means 19. Since the operation of the second camera shake correction unit 18 is exactly the same as that of the first embodiment,
Although the description is omitted, with such a configuration,
Since the field memories can be integrated into one and the final “cutout” uses the determined motion vector, the correction accuracy is sufficient.

In the above description of the embodiment, the correction gain calculation method of the flicker correction unit has been described as a process in which the corrected level has an average value, but the present invention is not limited to this. Further, the memory of the image stabilizing unit in the first and second embodiments of the second invention is a field memory, but a frame memory may be used. Also, the first to the third
Although the noise reduction unit in the embodiment of the invention has been described as a field cyclic type, it may be a frame cyclic type. The noise extraction processing characteristic has been described with reference to FIG. 14A as an example, but it is not limited to this. Further, although the case where the number of band divisions of the band division processing method is 2 is shown, the number is not limited to this.

[0062]

As described above, according to the first aspect of the present invention, the filter means for extracting a predetermined frequency from the input signal, the gain calculating means for calculating the gain for correcting the level fluctuation of the output of the filter means, and the input signal A level fluctuation correction unit configured by a multiplication unit that obtains a product of the gain calculation unit, a storage unit that stores a signal for a predetermined period, and a first subtraction unit that subtracts the storage unit output from the level fluctuation correction unit output. , A non-linear means for performing non-linear processing on the output of the first subtracting means, and a second subtracter for subtracting the non-linear means output from the output of the level fluctuation correction section and using the output as the input of the storage means. And a noise reducing section that outputs the noise reducing section as an output signal, the flicker can be removed from a signal accompanied by the flicker, and a sufficient S
The / N improvement effect can also be realized, and its practical effect is great.

A second aspect of the invention is a motion vector detecting means for detecting and outputting the direction and amount of the motion of the signal from the input video signal, a first storage means for storing the input video signal for a predetermined period, and a motion. First based on the output of the vector detection means
Image stabilization unit configured to output a part of the signal stored in the storage unit, a second storage unit that stores the signal for a predetermined period, and a second storage unit from the image stabilization unit output. First subtraction means for subtracting the means output, and
A non-linear means for performing non-linear processing on the output of the subtracting means, and a second subtractor for subtracting the non-linear means output from the shake correction unit output and using the output as the input of the second storage means.
And a noise reduction unit configured to use the output of the noise reduction unit as an output signal, the camera shake can be removed from a signal accompanied by the camera shake, and a sufficient S / N improvement effect can be realized. It is possible and its practical effect is great.

The third aspect of the present invention provides a motion vector prediction detection means for predicting and detecting the direction and amount of signal motion from an input video signal and outputting a prediction value and a detection value, and a motion vector prediction detection means output. An image sensor driving means for controlling driving of the image sensor based on the first memory means, a first memory means for storing the input video signal for a predetermined period, and a first memory means for storing the predicted output of the motion vector prediction detecting means. A first camera shake correction unit configured to output a part of the signal, a second storage unit that stores the signal for a predetermined period, and a second camera shake output from the first camera shake correction unit. First subtracting means for subtracting the output of the storing means, and a non-linear means for performing non-linear processing on the output of the first subtracting means,
The noise reduction unit configured by subtracting the output of the non-linear means from the output of the camera shake correction unit and using the output as the input of the second storage unit, the second storage unit, and the motion vector A second camera shake correction unit including a second control unit that controls so as to output a part of the signal stored in the second storage unit based on the detection output of the prediction detection unit. With the configuration in which the output of the second camera shake correction unit is used as the output signal, the camera shake can be removed from the signal accompanied by the camera shake, and a sufficient S / N improvement effect can be realized, and its practical effect is large.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a video signal processing device according to a first embodiment of the first invention.

FIG. 2 is a block diagram showing a configuration of a video signal processing device according to a second embodiment of the first invention.

FIG. 3 is a block diagram showing a configuration of a video signal processing device according to a first embodiment of the second invention.

FIG. 4 is a block diagram showing a configuration of a video signal processing device according to a second embodiment of the second invention.

FIG. 5 is a block diagram showing a configuration of a video signal processing device according to a third embodiment of the second invention.

FIG. 6 is a block diagram showing a configuration of a video signal processing device according to a fourth embodiment of the second invention.

FIG. 7 is a block diagram showing a configuration of a video signal processing device according to a first embodiment of the third invention.

FIG. 8 is a block diagram showing a configuration of a video signal processing device according to a second embodiment of the third invention.

FIG. 9 is a frequency distribution diagram for explaining the reason why flicker correction occurs.

FIG. 10 is a schematic diagram for explaining a state of flicker correction.

FIG. 11 is a schematic diagram showing the principle of electronic image stabilization.

FIG. 12 is a block diagram showing the configuration of a field recursive noise reduction circuit.

FIG. 13 is a block diagram showing the configuration of a field recursive noise reduction circuit.

FIG. 14 is a characteristic diagram showing the input / output characteristics of the noise extraction unit in the noise reduction circuit.

FIG. 15 is a block diagram showing a configuration of a band division type field cyclic noise reduction circuit.

[Explanation of symbols]

1 Flicker correction unit (level fluctuation correction unit) 2 BPF (filter means) 3 Gain calculation means 4 Multiplier (multiplication means) 5 Noise reduction section 6 First subtraction means 7 Second subtraction means 8 field memory (storage means) 9 Non-linear processing means 10 LPF (band splitting means) 11 HPF (band splitting means) 12 Synthetic means 13 Image stabilization unit 14 Motion vector detector 15 field memory (first storage means) 16 Memory control means (control means) 17 CCD drive pulse generator (imaging device drive means) 18 Second Image Stabilizer 19 Second memory control means (second control means) 22 line memory (first storage means) 23 field memory (second storage means) 24 First Memory Control Means (First Control Means) 25 motion vector predictor 26 Motion Vector Prediction Detector

Continuation of the front page (56) Reference JP-A-4-280171 (JP, A) JP-A-2-126777 (JP, A) JP-A-1-125068 (JP, A) JP-A-4-296176 (JP , A) JP 5-176218 (JP, A) JP 62-123880 (JP, A) JP 61-206378 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB) Name) H04N 5/21-5/217 H04N 5/225-5/247

Claims (8)

(57) [Claims] 1. A motion vector detecting means for detecting and outputting a direction and amount of a signal motion from an input video signal, a first storage means for storing the input video signal for a predetermined period, and an output of the motion vector detecting means. A camera shake correction unit configured to output a part of the signal stored in the first storage unit based on the above; a second storage unit that stores the signal for a predetermined period; First subtracting means for subtracting the output of the second storage means from the partial output; nonlinear means for performing nonlinear processing on the output of the first subtracting means; and subtracting the nonlinear means output from the shake correction section output And a second subtractor whose output is input to the second storage means, and a video signal processing device having the noise reduction unit output as an output signal. 2. The noise reduction unit includes: a second storage unit that stores a signal for a predetermined period; a first subtraction unit that subtracts the output of the second storage unit from an output of a camera shake correction unit; and the first subtraction. Band dividing means for dividing the output of the means into a plurality of bands; a plurality of non-linear means for performing non-linear processing on the plurality of outputs of the band dividing means; a combining means for combining the outputs of the plurality of non-linear means; Subtract the output of the synthesizing means from the output of the correction unit,
The video signal processing device according to claim 1, further comprising a second subtractor whose output is an input of the second storage means. 3. The image stabilization unit includes a motion vector predicting unit that predicts and outputs a direction and amount of a signal motion from an input video signal, and an image capturing unit that controls driving of an image sensor based on the output of the motion vector predicting unit. Element driving means, first storage means for storing the input video signal for a predetermined period, and control means for outputting a part of the signal stored in the first storage means based on the output of the motion vector prediction means The video signal processing device according to claim 1 or 2 , comprising: 4. A motion vector prediction detection unit that predicts and detects the direction and amount of signal motion from an input video signal and outputs a prediction value and a detection value; and an image sensor based on the output of the motion vector prediction detection unit. Image sensor driving means for controlling driving, first storage means for storing the input video signal for a predetermined period, and a signal stored in the first storage means based on the prediction output of the motion vector prediction detection means. A first camera shake correction unit configured to output a part of the first camera shake correction unit, a second storage unit that stores a signal for a predetermined period, and a second camera shake correction unit output from the first camera shake correction unit. First subtracting means for subtracting the output of the storing means, non-linear means for performing non-linear processing on the output of the first subtracting means, and subtracting the output of the non-linear means from the output of the first camera shake correction unit. A second subtractor whose output is the input of the second storage means; and a noise reduction section comprising: a second storage means; and the second storage means, based on the detection output of the motion vector prediction detection means, A second camera shake correction unit including a second control unit that controls so as to output a part of the signal stored in the second storage unit, and the second camera shake correction unit output. Video signal processing device for output signal. 5. The noise reduction unit includes: a second storage unit that stores the signal for a predetermined period; a first subtraction unit that subtracts the output of the second storage unit from the output of a first camera shake correction unit; A band dividing means for dividing the output of one subtracting means into a plurality of bands; a plurality of non-linear means for performing non-linear processing on the plurality of outputs of the band dividing means; and a synthesizing means for synthesizing the plurality of non-linear means outputs. A second subtraction of the output of the synthesizing means from the output of the first camera shake correction section, and the output thereof being the input of the second storage means,
5. The video signal processing device according to claim 4 , comprising: 6. The video signal processing device according to claim 1, wherein the storage period of the second storage means is one field or one frame. 7. The video signal processing device according to claim 1, wherein the storage period of the first storage means is one field or one frame. 8. The video signal processing device according to claim 3, wherein the storage period of the first storage means is n lines (n = 1,2,3, ...).