JP3419877B2 - Object recording device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [0001] BACKGROUND OF THE INVENTION 1. Field of the Invention
InobjectIn an image monitoring device for monitoringobject
of a bodyRecord imagesobjectIt relates to a recording device. [0002] 2. Description of the Related ArtobjectAs an example of a recording device,
For example, the one shown in FIG. 8 is known. This device is
Zoom in on a moving object in a surveillance area to
It is a device that records images. That is, for moving object detection
ITV camera (imaging means) 11
Is imaged and converted into an electric signal. From this camera 11
The image signal obtained is digitally converted by the A / D converter 12.
After being converted into data, the image memory 1
3 are sequentially stored. The moving object recognition unit 14
Recognize a moving object by referring to the contents of
As well as the position and size of the moving object. On the other hand, an ITV camera 15 for capturing detailed images
Has a zoom lens 16 and a swivel table 17,
The position information of the moving object obtained by the moving object recognition unit 14 is
The swivel base 17 is driven and controlled based on the zoom lens.
When the zoom 16 is zoomed up, the recognized movement is
Capturing moving objects in detail. Image signal from this camera 15
Is input to a VTR (video tape recorder) 18.
You. Therefore, by operating this VTR 18,
Images of recognized moving objects can be recorded in detail.
You. FIG. 9 is a flowchart of a conventional image recording operation.
Show Normally, images are captured from the camera 11 and monitored.
The moving object recognition unit 14 determines whether there is a moving object in the area.
Recognizes. In this case, the image data captured at a certain time
And compare it with previously captured image data
Thus, the presence or absence of a moving object is recognized. Specifically, two sheets
Calculates the difference between the digital data of the image pixel by pixel,
Pixels with a large difference can be regarded as changed pixels.
Where the changed pixel exists in a lump
Is considered to have some kind of changing object.
The presence of the body can be detected. [0005] In this way, a moving object is placed in the monitoring area.
If it is recognized that there was
And start recording process. Zoom up of a moving object
Move the camera 15 that captures the object in detail to the position of the moving object
And zoom lens 16
This can be realized by performing a step-up operation. For this purpose, the moving object recognition unit 14
This camera 15 is mounted on the basis of the provided position information.
The drive control of the swivel table 17 and the moving object recognition unit 1
The zoom lens 16 is moved based on the size information obtained in step 4.
Drive control to achieve an appropriate zoom ratio. [0007] The image recording process is performed by the camera 15.
Sends a recording start signal to the VTR 18 connected to
By setting TR18 to the recording state, the details of the moving object
Images can be recorded. This record is for moving object identification.
That the moving object has disappeared from the monitoring area
And continue until it is recognized. The moving object has disappeared
When it is recognized, a stop signal is given to the VTR 18.
Thereby, the recording operation of the detailed image is completed. [0008] The relationship between the elapsed time and the recorded image at this time is as follows.
As shown in FIG. The image at time ts−1 and the image at time ts
There is a change between. Therefore, the time ts-1
Capture images, recognize moving objects, move camera, zoom
The recording process starts when the system process is completed. Therefore,
Recording can be performed from the image at time ts + 1. The time te-1Has changed
Therefore, recording continues, and at the time te, the image changes.
Stop recording because it is gone. Therefore, the recorded image is
It is possible to record continuously 1 → 2 → 3 → 4. [0010] As described above, as described above,object
In order to detect
Since arithmetic processing is performed, it takes time for image processing. Also,
It takes time to start the VTR 18. For that, the camera
The detailed image captured at 15 shows the object perfectly perfectly
Not necessarily. In addition, the whole monitoring area andobjectTerritory where exists
Because I do not know the relationship with the area,objectIs all monitoring areas
I can't see where it appeared in. This
IsobjectThe same is true when disappears.
Where is the statue aloneobjectI do not know how disappeared. Accordingly, the present invention providesWhen recognizing an object,
Images can be recorded before the recognition, and objects can be
Recording, as well as the location of the object and
An object that can record its appearanceProvide a recording device
The purpose is to: [0013] An object recording apparatus according to the present invention.
Means for capturing an image in the monitoring area;
Image storage means for storing an image taken by the means,
By processing the image data in the image storage means,
Object recognizing means for recognizing an object in the monitoring area;
One image recorder that can digitally record image data
And when the object recognizing means recognizes the object,
From a predetermined time before theThe entire image is stored in the image storage
Read from the column, and the entire
While starting recording the body image, the object recognition means
When the object is recognized, it is recognized as time passes
The detailed image of the field of view of the enlarged object
And the object is recognized by the object recognizing means.
If it is no longer recognized, the entire image before and after that point is
By switching from the fine image, the field of view is the same size as before the object recognition
Detailed image at the same time and recorded as the entire image of
And the entire imageImage recording control means
I have it. [0014] [0015] [0016] [0017] [0018] [0019] Function of the present inventionobjectThe recording device isAn image
Recognize the object and record the image. After that, the example of an object
For example, zoom up and track an objectRecord detailed image of
You. Thus,Check the image of the object before recognition.
Indeed it can be recorded. Also an objectAppearance location and out
It is possible to record the current situation, the location of disappearance and the state of disappearance.
Wear. [0021] BRIEF DESCRIPTION OF THE DRAWINGS FIG.
Will be explained. FIG. 1 is a moving object recording apparatus according to the present embodiment.
It shows the configuration of FIG. This device is located within a certain monitoring area
Zoom in and track the moving object to
In this embodiment, a fine image is recorded.
There are two ITV cameras as steps and one is a moving object
Used for detection, the other has a zoom lens and swivel
It is used for capturing a detailed image of a moving object. Also, the image
Digital recording is performed using a magneto-optical disk as the recording medium.
Here is an example of the case. The same parts as those in FIG.
The description will be made with reference to FIG. In FIG. 1, two cameras 11, 15
Each image signal in the monitoring area input from the
After being converted into digital data by the D converters 12 and 19,
Each of the image data as image storage means is stored as image data.
The data is sequentially stored in the memory 13. The image memory 13 stores a predetermined file.
Images that have storage capacity for frames and are sequentially input
Is stored cyclically. The moving object recognizing unit 14 stores in the image memory 13
To refer to the stored input image from the camera 11
And recognize the moving object
For example, a process for obtaining the position and size of is performed. The camera 15 for capturing detailed images has a zoom
And a swivel table 17, which are
The drive is controlled by the controller 20. Camera control unit 2
0 is the position of the moving object determined by the moving object recognition unit 14
The information is converted to the angle of the swivel 17 of the camera 15 and the angle
Drive control of the swivel base 17 based on the degree information.
The size information of the moving object obtained by the moving object recognition unit 14
The zoom ratio of the zoom lens 16 is obtained based on the
By driving the zoom lens 16 according to the zoom ratio,
To a proper zoom ratio. The moving object recognition unit 14 recognizes a moving object.
If it is recognized, the images 11 and 15 are stored in the image memory 13.
The image including the moving object captured by
Therefore, these contents are sequentially transferred to the image recording / reproducing unit 21.
I do. The image recording / reproducing unit 21 transmits the transferred image data.
Add directory information and attribute information to
To the magneto-optical disk device 22 as an image recording means.
Thus, the information is sequentially recorded on a magneto-optical disk. Next, the recording operation of an image including a moving object will be described.
2 and the time chart shown in FIG.
This will be described with reference to a chart. In FIG. 3, 1 is
Indicates a monitoring area (imaging area), and 2 indicates a moving object.
You. First, the image signal from the camera 11 is A /
A / D conversion is performed by the D conversion unit 12, and the image memory 13 is successively converted.
Is input to The image memory 13 has a finite storage area.
If the last storage area has been used,
Take measures such as returning to the first storage area and storing
ing. That is, the image memory 13
Is stored cyclically. Image data stored in image memory 13
Is transferred to the moving object recognizing unit 14 and the moving object
And so on. In this case, there is no moving object in the image.
Then, the image previously taken into the image memory 13 and the next
Compare these as there are differences between
In this way, a moving object is recognized. More specifically, the difference operation between the images is performed in pixel units.
Calculation makes this difference in the moving object area.
Therefore, the difference pixels are larger than a certain value.
If so, it is determined that there is a moving object. Figure 3 shows the time chart.
The moment (time) ts of the appearance of this moving object
And store the frame number of the captured image as Mts.
Good. Naturally, the moving object is included in the image of frame number Mts-1.
Does not exist at all, and the background
Used as an image. For capturing a detailed image of the recognized moving object,
In order to move the moving object on the screen,
By converting the angle of the table 17 and moving the swivel table 17,
To move the field of view of the camera 15. Also, at the same time
Then, using the size information of the moving object, the zoom lens 16
By zooming in on a moving object,
The image is magnified to the size. In this way, a moving object captured in detail
The image signal is sent from the camera 15 to the A / D converter 19.
Here, the image data is converted into digital data and stored in the image memory 13.
Is stored. This image is referred to as frame number Bts + 1.
You. At the next instant ts + 2, the moving object moves to a different position.
Move. At this time, the moving object recognition unit 14
11 and the background image (Mts-1) described above.
And calculate the difference between the
By knowing the position, the moving distance and speed of this moving object
Degree can be determined. Therefore, the camera control unit 20 determines the current position.
Of the moving object after the calculation time from the
The position is estimated, and the turntable 17 is estimated based on the estimated position information.
Is controlled, and the field of view of the camera 15 is moved to that position.
You. The camera control unit 20 controls the size of the moving object.
From the size information, the appropriate zoom ratio of the zoom lens 16 is calculated.
Control the zoom lens 16 to obtain details of the moving object.
Capture the right image. This image has frame number Bts + 2 and
Become. In this operation, the moving object disappears from the monitoring area.
Continue until you do. Then, the moving object recognition unit 14
When the body disappears, the captured image and background image at that time
There is no difference between Frame the captured image at that time
The number is Mte. Then, the frame number Mte−1
The image shows the position of the moving object before it disappeared, and details at that time
Such an image has a frame number Bte-1. Here, is the moving object a frame number Mts?
From the image of Mte-1 and the frame number Bts
It exists in the image of Bte-1. If you record that image,
The moving object can be recorded, but the turntable of the camera 15
Assuming that a is a time until the moving object 17 follows the moving object, a
Moving object in image of frame number Mts + a to Mte-1
Body is detailed from frame number Bts + a to Bte-1
It is remembered. Therefore, both of these images are recorded.
Since the recording medium is wasted, the frame numbers Mts−1,
After recording Mts, ..., Mts + a and the whole image
, Frame numbers Bts + a, Bts + a + 1,.
Bte-1 and the detailed image are recorded, and finally the frame number
No. Mte-1, Mte, etc. to record the whole image
Can prevent unnecessary recording and save recording capacity
Thus, recording can be performed in a short time. In the example of FIG. 3, first, the image is recorded as an entire image.
After recording images 1 and 2, the detailed image was recorded as image 3,
4 and 5, and finally the whole image is recorded again.
7 is recorded. This image recording is performed by storing the image in the image memory 13.
By transferring the data to the image recording / playback unit 21,
Add directory information and attribute information to image data
To the magneto-optical disk drive 22,
This is performed by sequentially recording. It should be noted that in the magneto-optical disk drive 22,
When recording images, keep the time interval of the recorded images constant
Normally, for example, when a moving object exists for a long time
To save the recording capacity of the magneto-optical disk drive 22
Image to be recorded by the existence time of the moving object
Time interval is variable, and the number of records at one detection is constant
It is necessary to devise something like That is, the image memory
13 is not recorded continuously but skipped.
This is the method of recording. In this case, the moving object recognition unit 14
Measure the time from recognition of a moving object to its disappearance
By calculating the existence time of the moving object,
The time information is sent to the image recording / reproducing unit 21. And the picture
When recording an image in the image recording / reproducing unit 21,
Based on the existence time information of the moving object from the object recognition unit 14,
Calculate the appropriate number of recordable images (number of images)
While recording an image on the magneto-optical disk drive 22
It can be realized by doing. Hereinafter, a specific recording operation in this case will be described.
The flowchart shown in FIG. 4 and the tie shown in FIG.
This will be described with reference to a time chart. First, from camera 11
The A / D converter 12 converts the image signal of
The data is input to the image memory 13. The image memory 13
As described above, sequentially input images are stored cyclically. Image data stored in the image memory 13
Is transferred to the moving object recognizing unit 14 and the moving object
And so on. In this case, there is no moving object in the image.
Then, the image previously taken into the image memory 13 and the next
Compare these as there are differences between
In this way, a moving object is recognized. More specifically, the difference operation between the images in units of pixels is performed.
Calculation makes this difference in the moving object area.
Therefore, the difference pixels are larger than a certain value.
If so, it is determined that there is a moving object. Figure 5 shows the time chart.
The moment (time) ts of the appearance of this moving object
And store the frame number of the captured image as Mts.
Good. Naturally, the moving object is included in the image of frame number Mts-1.
Does not exist at all, and the background
Used as an image. For capturing a detailed image of the recognized moving object,
In order to move the moving object on the screen,
By converting the angle of the table 17 and moving the swivel table 17,
To move the field of view of the camera 15. Also, at the same time
Then, using the size information of the moving object, the zoom lens 16
By zooming in on a moving object,
The image is magnified to the size. Thus, the moving object captured in detail
The image signal is sent from the camera 15 to the A / D converter 19.
Here, the image data is converted into digital data and stored in the image memory 13.
Is stored. This image is referred to as frame number Bts + 1.
You. At the next instant ts + 2, the moving object moves to a different position.
Move. At this time, the moving object recognition unit 14
11 and the background image (Mts-1) described above.
And calculate the difference between the
By knowing the position, the moving distance and speed of this moving object
Degree can be determined. Therefore, the camera control unit 20 determines the current position.
Of the moving object after the calculation time from the
The position is estimated, and the turntable 17 is estimated based on the estimated position information.
Is controlled, and the field of view of the camera 15 is moved to that position.
You. The camera control unit 20 controls the size of the moving object.
From the size information, the appropriate zoom ratio of the zoom lens 16 is calculated.
Control the zoom lens 16 to obtain details of the moving object.
Capture the right image. This image has frame number Bts + 2 and
Become. In this operation, the moving object disappears from the monitoring area.
Continue until you do. Then, the moving object recognition unit 14
When the body disappears, the captured image and background image at that time
There is no difference between Frame the captured image at that time
The number is Mte. Then, the frame number Mte−1
The image shows the position of the moving object before it disappeared, and details at that time
Such an image has a frame number Bte-1. Here, is the moving object a frame number Mts?
From the image of Mte-1 and the frame number Bts
It exists in the image of Bte-1. If you record that image,
The moving object can be recorded, but the turntable of the camera 15
Assuming that a is a time until the moving object 17 follows the moving object, a
Moving object in image of frame number Mts + a to Mte-1
Body is detailed from frame number Bts + a to Bte-1
It is remembered. Therefore, as described above, both of these images are described.
Recording is a waste of the recording medium, so the frame number M
ts−1, Mts,..., Mts + a and the entire image
After recording, the frame numbers Bts + a, Bts + a + 1
 , ……, Bte-1 and detailed images are recorded.
Record whole images such as frame numbers Mte-1 and Mte
I do. This image is recorded in the image memory 1 as described above.
3 is transferred to the image recording / reproducing unit 21.
With this, directory information and attribute information such as
And sends it to the magneto-optical disk drive 22,
This is done by sequentially recording on a disc.
In some cases, images are recorded while thinning out the images as appropriate.
You. That is, the moving object recognition unit 14
Recognizing the disappearance of the body, it recognizes the moving object and then disappears
By measuring the time until the moving object
Interval, and the obtained existence time information is stored in the image recording / reproducing unit.
Send to 21. Then, the image recording / reproducing unit 21 outputs the image
When recording, the presence of a moving object from the moving object recognition unit 14
Calculates the appropriate number of recordable images (number of images) based on time information
Out, and thin out the image as appropriate in accordance with the calculated value.
The image is recorded on the physical disk device 22. Specifically,
For example, as the existence time of a moving object becomes longer,
Perform thinning processing such as increasing the number of images to be thinned.
And record the image. In the example shown in FIG. 5, the image is first recorded as an entire image.
After recording images 1 and 2, insert a detailed image
While recording, record images 3, 4, 5, and 6, and finally
The entire image is recorded as recorded images 7 and 8 again. In the above embodiment, the moving object is recognized.
Before moving objects disappear,
After storing in the image memory 13,
The case where an image is recorded on the storage device 22 has been described.
Ming is not limited to this, while capturing images
Recording may be performed simultaneously. Thus, the picture
By performing the recording process one after another while capturing the image,
More images can be recorded with less memory capacity
Noh. In this case, for example, the image memory 13
A storage capacity of 2 frames (2 planes) and 1 plane is written
When in the middle, the other side is configured to be read-only.
It can be realized by doing. Hereinafter, a specific recording operation in this case will be described.
6 and the tie shown in FIG.
This will be described with reference to a time chart. First, from camera 11
The A / D converter 12 converts the image signal of
The data is input to the image memory 13. The image memory 13
As described above, it has a storage capacity for two surfaces and one surface
When is writing, the other side is read-only
Is configured. Image data stored in the image memory 13
Is transferred to the moving object recognizing unit 14 and the moving object
And so on. In this case, there is no moving object in the image.
Then, the image previously taken into the image memory 13 and the next
Compare these as there are differences between
In this way, a moving object is recognized. More specifically, the difference operation between the images in pixel units is performed.
Calculation makes this difference in the moving object area.
Therefore, the difference pixels are larger than a certain value.
If so, it is determined that there is a moving object. Figure 7 shows the time chart.
The moment (time) ts of the appearance of this moving object
Is Mts. Of course,
No moving object exists in the image with the frame number Mts-1.
Used as a background image during subsequent recognition processing.
I have. The moving object recognizing section 14 recognizes a moving object.
First, the image (frame
The whole image one frame before the number Mts), that is,
That is, the image data of the frame number Mts-1 is stored in the image memory.
13 to the image recording / playback unit 21,
Sent to magneto-optical disk drive 22 for recording on magneto-optical disk
I do. To capture a detailed image of the recognized moving object
In order to move the moving object on the screen,
By converting the angle of the table 17 and moving the swivel table 17,
To move the field of view of the camera 15. Also, at the same time
Then, using the size information of the moving object, the zoom lens 16
By zooming in on a moving object,
The image is magnified to the size. In this manner, a moving object captured in detail
The image signal is sent from the camera 15 to the A / D converter 19.
Here, the image data is converted into digital data and stored in the image memory 13.
Is stored. This image is referred to as frame number Bts + 1.
You. At the next instant ts + 2, the moving object moves to a different position.
Move. The frame number Bts + 1 is stored in the image memory 13.
When the moving object recognition unit 14 captures the image of
1 frame more than the current image (frame number Bts + 1)
Image before the frame, that is, the image of the frame number Mts
Transfer data from the image memory 13 to the image recording / reproducing unit 21
By sending the light to the magneto-optical disk drive 22,
Record on a magnetic disk. At this time, the moving object recognition unit 14
11 and the background image (Mts-1) described above.
And calculate the difference between the
By knowing the position, the moving distance and speed of this moving object
Degree can be determined. Therefore, the camera control unit 20 determines the current position.
Of the moving object after the calculation time from the
The position is estimated, and the turntable 17 is estimated based on the estimated position information.
Is controlled, and the field of view of the camera 15 is moved to that position.
You. Further, the camera control unit 20 controls the size of the moving object.
From the size information, the appropriate zoom ratio of the zoom lens 16 is calculated.
Control the zoom lens 16 to obtain details of the moving object.
Capture the right image. This image has frame number Bts + 2 and
Become. The frame number Bts + 2 is stored in the image memory 13.
When the moving object recognition unit 14 captures the image of
1 frame more than the current image (frame number Bts + 2)
Image before frame, ie, frame number Bts + 1
Image data is transferred from the image memory 13 to the image recording / reproducing unit 21
To the magneto-optical disk drive 22
And records it on a magneto-optical disk. This operation is performed when the moving object is within the monitoring area.
Continue until it disappears. Then, the moving object recognition unit 14
Recognizes the disappearance of the moving object,
There is no difference from the background image. The captured image at that time
The frame number is Mte. Then, the frame number Mt
The image of e-1 shows the position of the moving object before it disappeared.
The detailed image at this time is frame number Bte-1. The moving object recognizing unit 14 determines that the moving object has disappeared.
Upon recognition, first, the image at that time (frame number Mt
e) The detailed image one frame before, that is, the frame
The image data of the program number Bte-1 is
By transferring the data to the image recording / reproducing unit 21, the magneto-optical
It is sent to the disk drive 22 and recorded on the magneto-optical disk. Record a detailed image of frame number Bte-1
After that, the image when the disappearance of the moving object is recognized
Overall image one frame before the image (frame number Mte)
Image, that is, image data of frame number Mte-1
Transfer from the image memory 13 to the image recording / reproducing unit 21
Is sent to the magneto-optical disk drive 22, and the
Record on the disc. Record the entire image of frame number Mte-1
After that, the whole picture when the moving object disappears
Image, that is, image data of frame number Mte
To transfer from the memory 13 to the image recording / reproducing unit 21
Is sent to the magneto-optical disk drive 22,
To record. Finally, for these continuous images,
Directory information and attribute information on a magneto-optical disk
By recording, the recording process of the image ends. FIG.
In the example of the first example, recording images 1 and 2
After that, detailed images are recorded as recorded images 3, 4, and 5.
Finally, the entire image is recorded again as recorded images 6 and 7.
You. [0074] As described in detail above, according to the present invention,object
Body image before its recognitionCan be recorded fromobject
bodyCan be recorded more reliably,objectof
Record appearance location and appearance, disappearance location and disappearance
can doobjectA recording device can be provided.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram schematically showing a configuration of a moving object recording device according to an embodiment of the present invention. FIG. 2 is a flowchart illustrating an image recording operation. FIG. 3 is a time chart illustrating an image recording operation. FIG. 4 is a flowchart illustrating an image recording operation according to another embodiment of the present invention. FIG. 5 is a time chart illustrating an image recording operation according to another embodiment of the present invention. FIG. 6 is a flowchart illustrating an image recording operation according to still another embodiment of the present invention. FIG. 7 is a time chart illustrating an image recording operation according to still another embodiment of the present invention. FIG. 8 is a block diagram schematically showing a configuration of a conventional moving object recording device. FIG. 9 is a flowchart illustrating an image recording operation in a conventional moving object recording device. FIG. 10 is a time chart illustrating an image recording operation in a conventional moving object recording device. [Description of Signs] 1 ... Monitoring area (imaging area), 2 ... Moving object, 11 ...
... ITV camera (imaging means) for detecting a moving object, 12,
19: A / D conversion unit, 13: Image memory (image storage unit), 14: Moving object recognition unit, 15: ITV camera (imaging unit) for capturing detailed images, 16: zoom lens, 17 turning table, 20 camera control unit, 21
Image recording / reproducing section 22, magneto-optical disk device (image recording means).

Continuation of the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) G08B 13/196 G01V 8/10 G11B 20/00 H04N 7/18

Claims (1)

(57) [Claims] 1. An image capturing means for capturing an image in a monitoring area, an image storing means for storing an image captured by the image capturing means, and image data in the image storing means. , An object recognizing means for recognizing an object in the monitoring area, one image recording means capable of digitally recording image data, and when the object recognizing means recognizes the object, the recognition time The whole image from a predetermined time before is read from the image storage means, recording of the whole image is started in the one image recording means, and when the object recognition means recognizes an object, A detailed image of the field of view obtained by sequentially enlarging the object recognized in accordance with the progress is recorded by switching from the whole image, and when the object recognizing means stops recognizing the object, before and after that time The entire image is switched from the detailed image and recorded as the entire image having the same size of the visual field as before the object recognition, and the detailed image at the same time and the entire image are recorded.
An image recording control device that records both images of an image.