JP3213348B2 - Pan / tilt processing circuit of camera shake correction camera - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pan / tilt processing circuit effective for a video camera having a camera shake correction function.

[0002]

2. Description of the Related Art A video camera having a camera shake correction function is
Detect motion vectors of the entire screen in the image formation area,
The image forming area in the imaging area is moved in accordance with the motion vector to correct the shake of the imaging screen due to camera shake. However, a motion vector is also generated by a pan or tilt operation. Therefore, the camera shake correction circuit also functions to prevent a pan operation and a tilt operation.

[0003] Therefore, for example, in the 1989 20th Conference on Image Engineering, a technique for permitting panning and tilting by limiting the amount of correction for movement during camera shake correction is proposed.

According to this technique, the relationship between the integral motion vector connecting the center of the imaging area and the center of the image forming area and the motion vector obtained by detecting the motion between fields in the image forming area is as follows. Is done.

[0005]

(Equation 3)

Where SV (n) is the X component (or Y component) of the integrated motion vector of the current field, and SV (n)
+1) indicates the X component (or Y component) of the integrated motion vector of the next field, V (n) indicates the X component (or Y component) of the motion vector of the current field, and K indicates a positive number of 1 or less.

In the above equation, if K is set to 1 or a value close to 1, camera shake correction is performed with good response, but it works to obstruct pan and tilt, and if K becomes smaller than 1, The camera shake correction becomes slower, but the centering process is performed without hindering the pan and tilt. Therefore, in this conventional technique, the optimum value of K is determined by experiments.

[0008]

However, the prior art described above is a numerical setting based on compromise, and it has never been possible to expect a sufficient effect in both motion compensation and pan-tilt measures. In particular, the camera shake correction has a remaining correction, and a sufficient correction cannot be performed.

Thus, it is necessary to switch and set K by detecting a pan state or a tilt state.

[0010]

The invention according to claim 1 is
A general motion vector of the subject in the image forming area is detected at a fixed period, correction data for changing the image forming area in the image forming area according to the motion vector is generated, and the image is generated based on the image output in the image forming area. In a camera shake correction camera that forms and derives a signal, a pan / tilt state identification unit that detects that the image forming area is in contact with the end of the imaging area for a predetermined period and detects a pan / tilt state, Following the detection of the pan / tilt state,
And a correction data generating means for generating correction data for forcibly moving the image forming area to the center of the image pickup area. It is a processing circuit. Claim 2
The invention according to the invention detects an overall motion vector of the subject in the image forming area at a fixed cycle, generates correction data for changing the image forming area in the imaging area according to the motion vector, and generates the correction data in the image forming area. In a camera shake correction camera that forms and derives a video signal based on an imaging output, a pan / tilt detecting a pan / tilt state by detecting that the video forming area is in contact with an end of the imaging area for a predetermined period.
Tilt state discriminating means, correction data for setting a value of K defined by the above equation (1) to be smaller than 1 after detecting the pan / tilt state and forcing the image forming area to the center of the imaging area. And generating the correction data when the image area forming means has moved to substantially the center of the imaging area or when it has been detected that the centering processing has been executed a predetermined number of times in the centering state of the correction means and the correction data generating means. And a centering end detecting means for canceling the center rig state of the means.

[0011]

According to the present invention, when the pan state or the tilt state is detected, the centering process is performed.
Becomes 1.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A video camera having a camera shake correction function employing the present invention will be described below with reference to FIG. still,
In the present embodiment, a case is described in which a motion vector has been generated a certain number of times in the same direction, and a pan or tilt state is detected, but the image forming area is in contact with the end of the imaging area for a certain period. This may be detected and regarded as a pan or tilt state.

First, an image pickup output of an image pickup area picked up by the image pickup device 1 is processed by the video processing means 2 and is output by
The data is digitized by the A conversion means 3 and stored in the frame memory 4. The converted data is supplied to a well-known motion vector detecting means 7, which compares the reference point of the previous field with the current field, and detects a motion vector connecting the position of the current field having the highest correlation with the video level of the reference point. Then, the motion correcting means 8 forms motion correction data according to the detected motion vector. This correction data is supplied to the motion vector detecting means 7 as reference point data of the next field, and to the memory control means 9 as data for selecting an image forming area. Therefore, the memory control means 9 selectively reads out only the necessary image forming area data from the stored imaging area data from the frame memory 4. The electronic zoom unit 5 interpolates the supplied data according to the zoom ratio, forms video data having a regular number of lines, and supplies the video data to the AD conversion unit 6. As a result, a video signal corrected for the camera shake correction is derived. In the configuration described above, the motion vector and the correction data are processed for each of the orthogonal X and Y components, and the vertical and horizontal corrections are made independently. In the configuration described above, the motion vector detecting means 7, the motion correcting means 8, and the memory control means 9 are constituted by microcomputer software.

The feature of this embodiment lies in the functional block of the motion compensating means 8 shown in detail in FIG. 1 and constituted by a microcomputer. First, the X component of the horizontal motion vector generated in each field is determined by the pan state detecting means 1.
It is input to 0X, and it is checked whether the same code is repeated 30 times. If it does not continue 30 times, it is regarded as a camera shake state, and the X-axis correction data generating means 12X is kept in a steady state. On the other hand, when 30 consecutive times, the X-axis correction data generating means 1
2X is switched to the centering state.

In this centering state, if the sign of the X component is inverted even once, the inversion detection output is generated by the pan state detection means, and the X-axis centering end detection means 11X
Is set to the operating state, and when the value of the X component becomes equal to or less than a predetermined value due to the centering, or when the end of the centering is detected when 60 fields have elapsed after the inversion detection state, the X-axis correction data generating means 12X is activated. Set to steady state.

The processing is similarly executed independently for the remaining Y component. Therefore, the tilt state detecting means 10Y
Is checked whether or not the same code is repeated 30 times. If it does not continue 30 times, it is regarded as a camera shake state and Y
The axis correction data generating means 12Y is kept in a steady state. vice versa,
If 30 consecutive times, the Y-axis correction data generation means 12X
Is switched to the centering state.

If the sign of the Y component is inverted even once in this centering state, an inversion detection output is generated by the pan state detection means, and the Y-axis centering end detection means 11Y
Is set to the operating state, and when the value of the Y component becomes equal to or less than a predetermined value due to the centering, or when the end of the centering is detected when 60 fields have elapsed after the inversion detection state, the Y-axis correction data generating means 12Y is turned on. Set to steady state.

Hereinafter, the correction data generating means 12X or 1
The specific function of 2Y will be described with reference to FIG.

The correction data of the correction data generating means is as follows:
In the steady state, both the X and Y components

[0020]

(Equation 4)

In the centering state,

[0022]

(Equation 5)

The following output is formed and derived.

Therefore, in the steady state, the constant switching means 13 sets K = 1 for the purpose of camera shake correction, and K = 0.98 because of centering until the centering end is detected after the pan or tilt state is identified. The multiplication means 14 multiplies the integral motion vector SV (n), which is the correction data of the previous field, by the selected K, adds the output K · SV (n) and the motion vector V (n) by the addition means 16, The output is generated as correction data.

Therefore, the correction data performs the correction following the camera shake state in the steady state, and follows the pan / tilt state in the centering state to perform the correction operation slowly to perform the centering.

[0026]

Thus, according to the present invention, the camera shake correction becomes slow at the time of panning / tilting, and natural panning / tilting is ensured. The photographing state is realized and the effect is great.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of a main part of the present invention.

FIG. 2 is a yesterday block diagram showing a further main part of FIG. 1;

FIG. 3 is an overall circuit block diagram of a video camera with a camera shake correction function that employs the present invention.

[Explanation of symbols]

 10X Pan state detecting means 11X X-axis centering end detecting means 12X X-axis correction data generating means 10Y Pan state detecting means 11Y Y-axis centering end detecting means 12Y Y-axis correction data generating means

──────────────────────────────────────────────────続 き Continuation of front page (56) References JP-A-2-246680 (JP, A) JP-A-3-216079 (JP, A) JP-A-2-75283 (JP, A) JP-A-61- 237581 (JP, A) JP-A-2-150180 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) H04N 5/222-5/257

Claims (2)

(57) [Claims] 1. An apparatus comprising: detecting an entire motion vector of a subject in an image forming area at a fixed period; generating correction data for changing the image forming area in an imaging area according to the motion vector; In a camera shake correction camera that forms and derives a video signal based on an imaging output, the image forming area is located at an end of the imaging area for a predetermined period,
Pan / tilt state identification means for detecting contact and detecting a pan / tilt state; and K defined by equation 1 in accordance with the detection of the pan / tilt state.
And a correction data generating means for generating correction data for forcibly moving the image forming area to the center of the imaging area. (Equation 1) 2. A method for detecting an entire motion vector of a subject in an image forming area at a constant period, generating correction data for changing the image forming area in an image pickup area according to the motion vector, and generating correction data in the image forming area. In a camera shake correction camera that forms and derives a video signal based on an imaging output, the image forming area is located at an end of the imaging area for a predetermined period,
A pan / tilt state discriminating means for detecting a touching state to detect a pan / tilt state; and a value of K defined by Expression 1 after detecting the pan / tilt state.
Is set to be smaller than 1, and the correction data generating means is in a centering state in which the image forming area is forcibly moved to the center of the imaging area. In the centering state of the correction data generating means, the image area forming means A centering end detecting means for canceling the center rig state of the correction data generating means when moving to a substantially center of the area or detecting that the centering processing has been executed a predetermined number of times; circuit. (Equation 2)