JPH05258184A - Security device - Google Patents

Abstract

PURPOSE:To automate security by discriminating whether an intruder exists within a security section or not based on information concerning difference between two different frame video signals inside a video output signal from a video camera. CONSTITUTION:A first memory 42 normally stores the video signal for one latest frame and the second memory 43 normally stores the video signal for the one frame immediately before the above frame. These video signals VSm and VSo are inputted to a difference arithmetic part 44 and an arithmetic processing for obtaining difference between them is executed. Then, video data which is indicated by an obtained difference video signal VSd is binarized and processed in a binarizing part 45 and binarized picture data VD is outputted. The data VD is inputted to an intrusion discriminating part 46 where it is checked whether binarized picture data is more than a prescribed data or not. When it is more than the prescribed one, a picture based on the difference video signal VSd is judged to be the one owing to the exsistance of the intruder within the security section.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a security device using a video camera.

[0002]

2. Description of the Related Art Conventionally, security camera devices have been widely used to guard a specific room or a specific area in a large room. Many of the conventional security camera devices have an endless video recording device that records time information together with the state of the required security area, and the image from the security camera device is monitored by security guards etc. If there is something intruding into the area, the guards will ring the crime prevention bell, report to the police, and the video recorded by the endless video recording device can be used as evidence.

[0003]

However, in this conventional security camera device, if there is an intrusion into the security area,
This cannot be detected automatically, and in order to detect an intruder, an intrusion detection device using infrared rays or ultrasonic waves must be installed. An object of the present invention is to provide a security device capable of electrically detecting the presence or absence of an intruder in a required security area by using a video signal from a video camera device.

[0004]

A feature of the present invention for solving the above-mentioned problems is that a video output from a video camera in a security device using a frame accumulation type CCD video camera or an image pickup tube type video camera. Processing means for performing a process for obtaining a difference between video signals of two different frames in the signal, intrusion determination means for responding to the processing means to determine whether or not the difference is larger than a predetermined amount, and the determination In response to the means, an output means for outputting an output for security when the difference is larger than the predetermined amount is provided. Further, when it is determined by the intrusion determining means that the difference is larger than the predetermined use, or irrespective of that,
Vector calculating means for calculating a movement vector of a moving body image in a captured image based on a pair of field video signals, and vector discrimination for discriminating whether or not the movement vector is directed to a predetermined specific protection target. Further comprising means, when the moving object is moving in a predetermined direction,
It can be configured to emit an output for security.

[0005]

The intrusion determining means determines whether or not there is an intruder in the security area based on the information on the difference between two different frame image signals in the image output signal from the video camera, which is obtained by the processing means. To be determined. Alternatively, the moving vector of the moving object is calculated by the vector calculating means, and an output for security is given when the moving vector points in a predetermined direction.

[0006]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described in detail below with reference to the drawings. FIG. 1 shows the configuration of an embodiment of a security device according to the present invention. The security device 1 according to the present invention has a video camera 2 installed in a room R and an endless video recorder 3 for recording a video output signal VS from the video camera 2. In the illustrated embodiment, the video camera 2 is a 2: 1 interlaced scanning type image pickup tube type video camera, and a predetermined security area including a safe K installed in a room R is within its field of view. The video output signal VS is adjusted so as to enter, and is also input to the image processing device 4 for performing image measurement by real-time processing.

The image processing apparatus 4 is constructed as a known microcomputer system, and its memory section 5 is provided.
The video output signal VS is processed according to the image processing program stored therein.

FIG. 2 is a functional diagram showing the contents of the processing of the video output signal VS executed in the image processing device 4. The memory unit 41 has a first memory 42 and a second memory 43, and has two consecutive 1's in the video output signal VS.
The frame video signals VSn and VSo are stored in the first and second memories 42 and 43, respectively, in the form of digital data. Here, the video signal VSn is a video signal for the latest one frame, and the video signal VSo is a video signal for one frame that is one frame before that. That is, the first memory 42 always stores the latest one-frame video signal, and the second memory 43 always stores the one-frame previous video signal for one frame. ..

These video signals VSn and VSo are input to the difference calculation section 44, and in this embodiment, a subtraction process is performed here to obtain the difference between the two. The video data indicated by the differential video signal VSd obtained as a result is 2
The binarization unit 45 binarizes the binarized image data VD by a known appropriate means. The binarized image data VD is input to the intrusion determination unit 46, where 2
It is checked whether the number of pixels of the binarized image data VD is a predetermined number or more. If the number of pixels is more than the specified number,
The image based on the differential video signal VSd is determined to be because the intruder is in the security area, and the first determination output H1 indicating that the intruder is in the security area is output.

On the other hand, the video output signal VS is stored in the memory section 41.
It is also input to the movement vector calculation unit 47 via.
The movement vector calculation unit 47 processes the image data of one field obtained by the video output signal VS and calculates the movement vector of the moving object included in the image data. Data processing for calculating the movement vector will be described with reference to FIG.

The video output signal VS is separated into an odd field video signal VS1 and an even field video signal VS2 in the signal separation section 51, and the signals VS1 and VS2 are input to the first calculation section 52 and the second calculation section 53, respectively. To be done.

The structure of the video output signal VS output from the video camera 2 will be described below. The image of one frame by the video output signal VS showing the state of the security area captured by the video camera 2 is obtained by combining one set of field images. Each of these field images is obtained as follows.

That is, when the security area captured by the video camera 2 is scanned by the interlace system, the security area is first scanned by scanning in the odd-numbered scanning lines.
An odd field video signal is obtained. In this case, since there is no shutter, if the intruder is moving in the guarded area, the above scanning is performed during the exercise period. Then, the security area is scanned in the same field of view by the scanning according to the even-numbered scanning lines, and the even-field video signal is obtained.

As shown in FIG. 4A, this interlaced scanning is performed in synchronization with each pulse P1, P2, ... Of the vertical synchronizing signal output every 1/60 second, in an odd field. Suppose that a scan of 1 is performed in time slot T1 of 1/30 second duration, the corresponding even field scan is also performed in time slot T2 of 1/30 second duration. Therefore, in this case, the time slots T1 and T2 overlap by 1/60 seconds.

The odd field video signal thus obtained in the time slot T1 is applied to the time slot Ta of the video signal VS shown in FIG. 4B. On the other hand, the even field video signal obtained in the time slot T2 is applied to the time slot Tb of the video signal VS shown in FIG. 4B, and the television video signal VS is produced in this manner.

The odd field video signal VS1 and the even field video signal VS2 are image data D1 and D2 showing the state of the guard area for one frame according to the respective signals VS1 and VS2 by the first operation unit 52 and the second operation unit 53. Are obtained respectively.

The image data D1 is, for example, as shown in FIG. 5 (a), every other scanning line, ...
Based on the odd field video signal VS1 having the image information of, the contents of the scanning lines, ... Are interpolated using the contents of the scanning lines ,. Is created as data indicating the image information of one frame.

On the other hand, the image data D2 corresponds to one frame by the same interpolation as that shown in FIG. 5B based on the image information of the scanning lines included in the even field video signal VS2. It is created as data indicating the image information of.

Therefore, if there is any moving body such as an intruder moving in the security area, the moving body image x according to the image data D1 based on the image information of the odd number field.
And image data D2 based on the image information of the even field
The moving body image y according to is, as schematically shown in FIG.
There will be some deviation along the movement direction, and some of them will overlap.

The set of image data D1 and D2 thus obtained is given to the logical product calculating section 54, where each image according to one image data D1 and the other image data D1.
An AND operation is executed with each image according to 2. Then, combination data CB indicating a combination of images in which a logical product image according to the logical product calculation result exists is provided to the movement vector calculation unit 55.

The movement vector calculation section 55 is further provided with image data D1 and D2, and the barycentric position of the image is calculated based on these. Further, the movement vector of the position of the center of gravity in each image pair is calculated based on the combination data CB indicating the combination of images, and the movement vector data VR indicating this result is output.

Returning to FIG. 2, the movement vector data VR is given to the movement vector discriminating section 48. If there is no moving object in the security area, the values of the obtained movement vector are all zero, but if there is a moving object, movement vector data having some value will be obtained,
The movement vector discriminating unit 48 discriminates whether or not each direction of the movement vector having any value is toward the safe K. If it is determined that the direction of even one of the obtained movement vectors is toward the safe K, the second determination output H2 is output.

The output section 49 outputs the first and second discrimination outputs H.
Output signal O for security only when the first and second discrimination outputs H1 and H2 are simultaneously output in response to 1 and H2.
Is output. Note that, here, the movement vector determination unit 48
Is configured to determine whether or not the movement vector is heading to the safe K, but when there are a plurality of security objects in the security area, the movement vector is directed to each security object. Alternatively, the second determination output H2 may be output when the movement vector is directed to at least one of the security targets.

As shown in FIG. 1, the output signal O is input to the security device 5 as an output of the image processing device 4 and operates the security device 5. That is, by processing the image of the security area captured by the video camera 2 based on the video output signal VS from the video camera 2 as described above, the presence of a moving object in the security area, and the moving object When it is determined that the vehicle is moving toward K, the security device 5 is activated.

However, the security device of the present invention is not limited to the configuration of this embodiment, and outputs the output signal O only when the intrusion determination unit 46 determines that a moving object exists in the security area. Alternatively, the moving vector discriminating unit 48 may output the output signal O only when the moving object in the guarded area faces the safe K, which is a predetermined guarded object. .. Further, the security device 5 can be an emergency bell, a police report device, or any other known appropriate device.

Since the state of the security area is recorded by the endless video recorder 3 in parallel with the above-mentioned processing operation in the image processing apparatus 4, it goes without saying that the recorded image is useful for criminal investigation.

FIG. 7 is a flow chart showing an image processing program executed by the image processing apparatus 4. By executing the image processing program 60, the image processing according to the functional diagram shown in FIG. 2 already described is executed.

When the execution of the image processing program 60 is started, first, in step 61, the video data corresponding to the video signals of two consecutive frames of the video output signal VS are input, and the video signals of these two frames are input. In step 62, difference calculation processing is performed to obtain a difference between the image of one of the frame video signals and the image of the other frame video signal.

At the next step 63, the difference image data obtained at step 62 is binarized. For this binarization, for example, the binarization method according to Otsu's method for automatically determining the threshold value for binarization may be adopted. Two
After the binarization, step 64 is entered, where it is judged whether or not the number of pixels of the binarized image data is larger than a predetermined value M.
If the number of pixels of the binarized image data is less than or equal to the predetermined value M, it is determined that the difference image data does not represent an intruder, and the determination result of step 64 is NO, and step 6
Return to 1. On the other hand, the number of pixels of the binarized image data is a predetermined value M
If it is larger, it is determined that the difference image data represents an intruder, the determination result of step 64 is YES, and the process proceeds to step 65.

In step 65, based on the video data input in step 61, the movement vector of the image of the moving body based on the video data is calculated.

This movement vector calculation processing will be described in detail with reference to FIG. When the movement vector calculation is started, first, in step 71, the odd field video signal VS1 is generated based on the video output signal VS.
And the even field video signal VS2 are separated,
Image data D according to these signals in step 72
1 and D2 are created respectively.

In the following description, the image represented by the image data D1 is referred to as the image at the first time, and the image represented by the image data D2 is referred to as the image at the second time.

Next, in step 73, data processing for binarizing each of the image data D1 and D2 is performed. The data processing for binarization is performed by converting each data D1 and D2 into binary image data by using a threshold value set appropriately. As a result, binarized data Da and Db indicating the state of the binarized security area are obtained.

In the next step 74, unnecessary information removal, that is, noise removal processing is executed based on these binarized data Da and Db. This process is the noise removal process of the image periphery by expansion and contraction. A process of removing an image having a number of pixels equal to or less than a preset number of pixels as noise. Is included. However, it goes without saying that other appropriate noise removal processing may be performed in addition to these or in place of some of them.

Next, in step 75, step 7
Labeling is performed on each image in each of the images at the first time and the second time remaining as a result of the noise removal processing of No. 4, and the center of gravity and the area are calculated for all the labeled images.

After that, the process proceeds to step 76, where a logical product is calculated between each image of the image at the first time and each image of the image at the second time. Then, in step 77,
As a result of the logical product operation, combination data having a value and showing a combination of the image label at the first time and the image label at the second time is obtained.

For example, as shown in FIG. 9, when the image of the label A1 and the image of the label A2 have a correspondence relationship and the positional relationship shown in FIG. 9, the result of the logical product of these images is as shown in FIG. When the area is larger than a predetermined value, it means that both images are moving object images formed by the same moving object. Therefore, in this case, the combined data showing one set of images indicated by the labels A1 and A2 is obtained. This operation is performed for all combinations to determine the image combination.

In step 78, in order to determine whether or not each combination of the images obtained as described above is valid, it is checked whether or not the following conditions are satisfied. (A) Each image is combined with only one of the images at the other time. (B) The difference between the areas of one set of images is within 20%. (C) The straight line connecting the centroids of the pair of images passes through the logical product image obtained by these. Only the combination data that satisfies each of the conditions (a), (b), and (c) is captured as the correct combination data, and any combination that does not satisfy any one of the conditions (a) to (c). Matching data is removed.

In step 79, the center of gravity W1 of the image A1 at the first time is selected from the pair of images shown by the combination data of which the validity is recognized, as shown in FIG. 9, for example.
From this, a vector VC corresponding to the barycenter position W2 of the image A2 at the second time is calculated, and this vector VC is set as the movement vector of the moving body. In this way, the movement vector is calculated as described above for the image pairs indicated by all the combined data. As a result, each movement vector of the moving body in the image becomes clear.

Returning to FIG. 7, in step 66, based on the movement vector calculated in step 65, it is determined whether or not there is a moving body moving in a predetermined direction, that is, a direction toward the safe K, and the determination is made. If the result is YES, the routine proceeds to step 67, where the output signal O is output and the alarm device 5 is activated. On the other hand, if it is determined in step 66 that there is no moving object moving in the direction toward the safe K, the process returns to step 65 and the movement vector is calculated again.

As described above, the presence / absence of an intruder is determined based on the difference image data obtained by the difference calculation process. When it is determined that there is an intruder, the movement vector of the intruder is further calculated, When the direction of the movement vector is the predetermined direction, it is determined that someone is approaching the security target, and the alarm device is activated.

[0042]

According to the present invention, as described above, by processing the image data obtained from the video camera for monitoring the security area, it is possible to determine whether or not an intruder is present in the security area without the eyes of the security officer. Since it can be determined whether or not the security is automated. It is also possible to calculate a movement vector of a moving object within the guarded area and issue a warning only when the movement vector is directed in a predetermined direction. It is also possible to output an alarm only when someone is approaching a security target in a security area where access is permitted. In any case, it is possible to detect intruders by electrically processing the output from the video camera, so it is possible to reduce the personnel for security, and the cost of security is greatly reduced. Can be made

[Brief description of drawings]

FIG. 1 is a configuration diagram of an embodiment of a security device according to the present invention.

FIG. 2 is a functional diagram showing the configuration of the image processing apparatus shown in FIG.

FIG. 3 is a detailed functional diagram showing the configuration of a movement vector calculation unit in FIG. 2 in detail.

FIG. 4 is an explanatory diagram for explaining a configuration of an interlaced television video signal.

FIG. 5 is an explanatory diagram for explaining a method of creating image data for one frame by interpolation based on an odd field video signal.

FIG. 6 is a diagram schematically showing how a moving body image is displaced according to a pair of image data.

7 is a flowchart showing an image processing program executed in the image processing apparatus of FIG.

FIG. 8 is a detailed flowchart of steps of calculating a movement vector in FIG.

9 is a diagram for explaining calculation of a movement vector executed in the image processing apparatus of FIG.

[Explanation of symbols]

 1 Security Device 2 Video Camera 4 Image Processing Device 44 Difference Calculation Unit 46 Intrusion Discrimination Unit 47 Movement Vector Calculation Unit 48 Movement Vector Discrimination Unit 49 Output Unit VS Video Output Signal VSn, VSo Video Signal for One Frame R Room

Claims (2)

[Claims] 1. A security device using a frame-accumulation type CCD video camera or an image pickup tube type video camera, for obtaining a difference between video signals of two different frames in a video signal output from the video camera. Processing means for performing processing, intrusion determination means for responding to the processing means to determine whether or not the difference is larger than a predetermined amount, and security for responding to the determination means when the difference is larger than the predetermined amount. And a means for giving the output of the security device. 2. A security device using a frame-accumulation CCD video camera or an image pickup tube type video camera, wherein a moving object within a picked-up image based on a pair of field video signals among video signals output from the video camera. Vector calculation means for calculating the movement vector of the image, vector discrimination means for discriminating whether or not the movement vector is directed to a predetermined specific protection target, and the movement vector in response to the vector discrimination means A security device comprising means for issuing an output for security when the security device faces the specific protection target.