JP2531893B2 - TV image subject movement detection method - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention facilitates monitoring by a television camera or identifies from an X-ray television camera by identifying whether or not a subject image obtained by the television camera is moving. The present invention relates to a TV image subject movement detection method used for automating image processing of an obtained television signal.

[0002]

2. Description of the Related Art FIG. 6 is a block diagram showing the configuration of a conventional TV image subject movement detection method. The TV camera 1 generates an analog television signal 11 from an optical image of a subject. The A / D converter 2 converts an analog television signal into a digital television signal 12. The one-screen memory 3 stores the digital television signal 12 in one screen (1
Frame), and the delayed television signal 13 is read out sequentially when one frame has elapsed from the input.
Output as The subtractor 5 generates a signal 15 which is the difference between the two digital television signals 12 and 13 having a time difference of one frame. The D / A converter 4 converts the digital delayed television signal 13 into an analog television signal 14 and sends it to the TV monitor.

Now, as shown in FIG.
It is assumed that the user moves from A to B in the frame period. And A
It is assumed that the subject image of the person located at the point is in the frame represented by the television signal 13, and the subject image of the person located at point B is in the frame represented by the television signal 12. At this time, the amplitude of the difference signal 15 has a larger value at the positions A and B than in the case where the subject image does not move. Even if there is no movement in the subject image, the amplitude of the difference signal does not become zero. This is because the television signal obtained from the TV camera 1 contains noise. Even if the television signal 11 obtained from the TV camera 1 has some noise, if there is a sufficient level difference between the signal of the moving subject image and the noise, that is, if the S / N is sufficiently large, By comparing the amplitude of the difference signal 15 with a predetermined threshold value, it is possible to detect whether or not the subject image has moved.

[0004]

When a subject is imaged and acquired under sufficient brightness, there is a large level difference between the subject image and noise, so the movement of the subject image can be detected by the circuit of FIG. However, in a video signal obtained by imaging in a dim environment, the difference in level between the signal of the subject image and noise is small, and S
Since the / N ratio is small, it is often the case that the object image cannot be detected even if the difference signal 15 is compared with the threshold value. Since it is generally impossible to expect sufficient illumination when a TV camera is used to monitor the intrusion of people, the circuit of FIG.

An image of a subject such as a human body is obtained by an X-ray TV camera, a television signal is processed only when the subject is stationary to obtain a clear image, and when the subject is moving, the television is displayed. An apparatus that detects movement of a subject image is required in order to prevent signal processing because the image is blurred even if the signal is processed. However, since the output of the X-ray TV camera contains a higher level of noise than that of the optical TV camera, the circuit of FIG. 6 cannot reliably detect the movement of the subject image. Therefore, in the past, a subject image was displayed on a TV monitor and the operator
While watching the V monitor, the transmission path of the television signal is manually switched depending on whether the subject image is stationary or moving, and the television signal is passed through the image processing device or the image processing device. Was bypassed.

Therefore, an object of the present invention is to provide a TV image capable of detecting the movement of the object image represented by the television signal even if the noise level in the television signal is high and the S / N ratio cannot be sufficiently obtained. The object is to provide a method for detecting movement of a subject image.

[0007]

In order to solve the above-mentioned problems, the present invention provides the following means. A TV image subject movement detection method for detecting a movement of a subject image of an image represented by the television signal by imaging a subject with a television camera and generating a television signal, which is a digital television. A means for generating a signal, a delay means for giving the digital television signal a time delay which is an integral multiple of one frame period, a digital television signal which is an input and an output of the delay means, and a signal of a difference between these signals. The subtraction means for generating, and the histogram generation means for generating a histogram relating to the difference signal for at least one frame,
The histogram generation unit stores in advance a histogram of the difference signal for one frame related to the television signal when a still object is captured by the television camera as a still image histogram, and the difference signal of the input frame is input. The histogram obtained by subtracting the still image histogram from the histogram of the above is output as data representing the movement of the subject image.
V image subject movement detection method. A TV image subject movement detection method for detecting a movement of a subject image of an image represented by the television signal by imaging a subject with a television camera and generating a television signal, which is a digital television. A means for generating a signal, a delay means for giving the digital television signal a time delay which is an integral multiple of one frame period, a digital television signal which is an input and an output of the delay means, and a signal of a difference between these signals. The subtraction means for generating, and the histogram generation means for generating a histogram relating to the difference signal for at least one frame,
The histogram generation unit stores in advance a histogram of the difference signal for one frame related to the television signal when a still object is captured by the television camera as a still image histogram, and the difference signal of the input frame is input. The area of the positive portion of the histogram obtained by subtracting the still image histogram from the histogram for is obtained by integration, and indicates that the subject image has moved when the outer area exceeds a predetermined threshold value. TV image subject movement detection method.

[0008]

In the present invention, the difference signal obtained by the subtracting means is guided to the histogram generating means, and the histogram generating means determines the movement of the subject image. The histogram is (his
This is a statistical term, and the abscissa indicates a section obtained by dividing the observation value region, and the height indicates the number of observations (the number of hits) that appear in each section. FIG. 3 is an example of a histogram of a digital television signal. In this figure, it is assumed that the density of each pixel of the television signal is represented by 8 bits, and each pixel of the television signal is 0 for density.
It is divided into 256 stages up to 255, and 1 at each concentration
The number of pixels in the frame is plotted on the height and drawn on a graph. Pure black has a density of 0 (all bits 0), pure white has a density of 25
5 (all 1s), gray is 128 (10000000)
It is represented by.

The histogram generating means creates a histogram for the difference signal output from the subtracting means. In the histogram of the difference signal, the horizontal axis represents the magnitude of the difference signal, that is, the difference in density between the input and output television signals of the delay means as in FIG. 3, and the number of pixels belonging to each density difference is plotted in the height direction. . 4A and 4B show examples of histograms of the difference signals. FIG. 3A shows a case where the subject image represented by the television signal does not move at all, that is, a still image. At this time, since the input and output of the delay means represent the same subject image, if there is no noise, the histogram is one vertical line standing at the position of 0 on the horizontal axis (the height of the vertical line is within one frame). Number of pixels). However, in reality, the television signal contains not only the image signal of the subject image but also noise of a considerable level, so that the histogram has a distribution curve spread in the horizontal axis direction as shown in FIG. Become.

FIG. 5 is a diagram exemplifying images represented by the input (B) and output (A) of the delay means and the output (C) of the subtraction means when the subject image moves between frames. If the subject image P shown in FIG. 5A moves while being delayed by the delay means to become the subject image Q shown in FIG. 5B, the output of the subtracting means is the output of FIG. The subject images P ′ and Q ′ are as shown in FIG. In FIG. 5, it is assumed that the hatched subject images P, Q, P ′ are pure black, the outlined subject image Q ′ is pure white, and the background is gray. The subtraction means moves to the black direction when the difference between the two television signals is + (plus), to the white direction when the difference is − (minus), and to the gray between the true white and the true black when the difference is 0. A difference signal in which the density of each pixel is set is generated.

FIG. 4B shows a histogram of the difference signal represented by the image of FIG. 5C. As shown in the figure, when the subject image moves between the two frames input to the subtraction means, a hump that is substantially symmetrical about the 0 position is generated in the histogram of the difference signal, and the difference in the case of a still image is generated. This is very different from the signal histogram (FIG. 4A).
Therefore, FIG. 4 showing the distribution of noise components in a still image.
FIG. 4 regarding a moving subject image is stored in advance as the still image histogram of the histogram a in FIG.
The still image histogram a is subtracted from the histogram b in (B) to generate the histogram c in FIG. 4C, and the movement of the subject image is detected from the data in the histogram c. For example, it may be determined that the subject image has moved when the area of the area (the portion shown by hatching) where the number of pixels is positive in the histogram c exceeds a predetermined threshold value.

In the present invention, the influence of noise on the television signal is largely removed by the process of subtracting the histogram a from the histogram b in FIG. 4, so that even if the television signal contains a considerable level of noise, the object Image movement can be detected with high reliability.

Next, the present invention will be described in more detail with reference to examples.

【Example】

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention. The TV image subject movement detection method of the present embodiment includes a histogram generator 6 in addition to the conventional method of FIG. The A / D converter 2 is an analog television signal 11
Is converted into an 8-bit digital television signal 12. The one-screen memory 3 stores an 8-bit digital television signal 12 for one screen (one frame) and outputs a digital video signal 13 delayed by one frame from the input. The subtractor 5 generates a difference (density difference) between the digital video signals 12 and 13 having a time difference of only one frame period for each pixel, and outputs a difference signal 15.

The histogram generator 6 includes a commercially available histogram generation IC (for example, IP90C0 manufactured by Sumitomo Metal Co., Ltd.).
1) is provided, and this histogram generation IC
A histogram of the difference signal 15 for the frame is generated. When the subject images in the television signals 13 and 12 are exactly the same, that is, when the relative position between the subject and the TV camera 1 does not change within one frame period, the histogram of the difference signal 15 for one frame is as shown in FIG. become. If the television signals 12 and 13 have no noise component, the histogram of the difference signal 15 will be one line standing at the position of the horizontal axis 0. However, in reality, since the noise mainly generated in the TV camera 1 is included in the television signals 12 and 13 as it is, the difference between the densities represented by the two signals often does not become zero in a pixel with noise. Therefore, the difference signal 1
The density differences of both signals shown by 5 are symmetrically distributed in the + direction (right side) and the-direction (left side) about the horizontal axis 0, and the histogram of the difference signal 15 is as shown in FIG. 4 (A). The histogram generator 6 obtains the histogram of FIG. 4A determined by the characteristics of the TV camera 1 and stores it in advance as a still image histogram.

Now, the television signal 13 for one frame including the subject image P shown in FIG.
If one frame of the television signal 12 including the subject image Q shown in (B) is input to the subtractor 5,
The image represented by the difference signal 15 for one frame output from the subtractor 5 is as shown in FIG. In FIG. 5, the subject images P, Q, P'are black, and the subject image Q '
Is pure white. In the background of FIGS. 5A and 5B, the TV
The noise mixed in the video signal by the camera 1 is shown. When the difference signal 15 is +, it is represented by the intermediate color in the white direction, and when the difference signal 15 is-, it is represented by the intermediate color in the black direction. When expressed as an image, a histogram of the difference signal for one frame, which is as shown in FIG. 5C, is shown in FIG. 4B. As shown in this figure, when the subject image moves between two frames, the densities in the difference signal 15 of the television signals of these two frames are centered on the horizontal axis 0 in the + direction (right side) and − direction (left side). And the histogram of the difference signal 15 for one frame is as shown in FIG.

The histogram generator 6 subtracts the histogram a of FIG. 4 (A) from the histogram b of FIG. 4 (B) to generate a histogram c shown in FIG. 4 (C), and a positive portion of the histogram c (see FIG. 4). The area of (C) is shown by hatching), and the movement detection signal 16
Output as Since the histogram a in FIG. 4A represents the distribution of the noise component, the histogram c in FIG. 4C obtained by subtracting the histogram a from the histogram b in FIG. It represents the subject image component. However, in the histogram c, the negative region in the Y-axis direction corresponds to the noise component, so it is desirable to detect the presence or absence of the moving subject image without taking this region into consideration.

The movement detection signal 16 is approximately proportional to the amount of the area moved between frames in the subject image. Therefore, in a device intended to monitor the presence or absence of an intruder with a TV camera, it can be determined that an intruder is present when the movement detection signal 16 exceeds a predetermined threshold value.

There is a technique of performing a process called integration on a television signal output from an X-ray TV camera to generate a clear X-ray image. However, in image processing that handles a television signal of an X-ray image, when the subject image moves, the subject image becomes rather unclear due to integration. Therefore, the television signal 14 is passed through the image processing apparatus only when the subject image is stationary, and the television signal 14 is caused to bypass the image processing apparatus when the subject image is moving. Using the embodiment of FIG. 1, the movement detection signal 16
When the level of the movement detection signal 16 is equal to or lower than the predetermined value, it is considered that the subject image is moving and the television signal 14 is automatically bypassed to the image processing device. The television signal 14 can be automatically passed through the image processor, assuming that the image is stationary.

In the embodiment shown in FIG. 1, as described above, the histogram generator 6 performs the process of removing the influence of the noise component contained in the difference signal 15. So TV
Analog television signal 1 output from camera 1
Even if the noise level in 1 is high and the S / N ratio is small, adoption of this embodiment makes it possible to detect with high reliability whether or not the subject image has moved. Therefore, the present embodiment is applied to a TV camera monitoring system that takes an image with dark illumination, or a system that automatically performs a switching operation to pass or bypass the output of the X-ray TV camera to the signal processing device. The reliability of can be greatly improved.

In addition, in the histogram generator 6, as shown in FIG.
The area of the area shown by hatching in (C) is output as the movement detection signal 16. However, the histogram generation means of the present invention is not limited to the method of this embodiment, and for example, compares the height of the histogram c in FIG. Corresponding to the number of hits in a random histogram) exceeds the threshold, the movement detection signal 16
Is set to a high potential (High Level), and the movement detection signal 16 is set to a low potential (Low L) for other frames.
There is no problem in adopting the method of "Ev.

[0022]

As described above in detail with reference to the embodiments, according to the present invention, even if the noise level in a television signal is high and the S / N ratio is small, the object represented by the television signal is TV that can detect image movement
An image subject image movement detection method can be provided.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.

FIG. 2 is a conceptual diagram showing movement of a subject image displayed on a display screen of a television monitor.

FIG. 3 is a diagram illustrating a histogram relating to an 8-bit digital television signal for one frame.

FIG. 4 is a diagram illustrating a histogram generated by a histogram generator 6 in FIG.

5 is a diagram exemplifying an image for one frame by a difference signal generated by a subtractor 5 in FIG. 1 when a subject image moves between frames.

FIG. 6 is a block diagram showing a configuration of a conventional TV image subject image movement detection method.

[Explanation of Codes] 1 TV camera 2 A / D converter 3 1 screen memory 4 D / A converter 5 Subtractor 6 Histogram generator 11 Analog television signal 12 Digital television signal 13 Delayed television signal 14 Analog television Signal 15 Difference signal 16 Movement detection signal

Claims (2)

(57) [Claims] 1. A TV image subject movement detection method for detecting a movement of a subject image of an image represented by the television signal by imaging a subject with a television camera to generate a television signal. A means for generating a digital television signal, a delay means for giving the digital television signal a time delay which is an integral multiple of one frame period, and a digital television signal for input and output of the delay means. And subtraction means for generating a signal of the difference signal, and histogram generation means for generating a histogram relating to the difference signal for at least one frame, wherein the histogram generation means is used when a still object is imaged by the television camera. A histogram of the difference signal for one frame relating to the television signal is displayed as a still image hist. Prestores as grams, T, characterized in that for outputting a histogram obtained by subtracting the static image histogram from the histogram of the difference signal of the frame which is input as data representing the movement of the object image
V image subject movement detection method. 2. A TV image subject movement detection method for detecting a movement of a subject image of an image represented by the television signal by imaging a subject with a television camera to generate a television signal. A means for generating a digital television signal, a delay means for giving the digital television signal a time delay which is an integral multiple of one frame period, and a digital television signal for input and output of the delay means. And subtraction means for generating a signal of the difference signal, and histogram generation means for generating a histogram relating to the difference signal for at least one frame, wherein the histogram generation means is used when a still object is imaged by the television camera. A histogram of the difference signal for one frame relating to the television signal is displayed as a still image hist. The area of the positive portion of the histogram, which is stored in advance as a gram and is obtained by subtracting the still image histogram from the histogram related to the difference signal of the input frame, is obtained by integration, and the area exceeds a predetermined threshold value. A TV image subject movement detection method characterized in that the subject image is sometimes moved.