JP3219582B2 - Object monitoring device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an object monitoring apparatus for automatically monitoring whether a predetermined moving object outdoors has stopped at a predetermined location by image processing. For example, there are many objects other than specific moving objects outdoors, such as monitoring the parking situation in an outdoor parking lot with many streets and monitoring the entry and exit of airplanes in an airport apron with many work vehicles. The present invention relates to an object monitoring apparatus in a place where the object is monitored.

[0002]

2. Description of the Related Art A conventional object monitoring apparatus will be described with reference to FIG. In FIG. 8, first, a TV camera 65 is installed at a place to be monitored, and an area to be monitored by the user in an image input from the TV camera 65 is determined.

In an image I1... In in which no moving object exists in the monitoring area defined by the user, an image I1.
Accumulate only once. Then, an image 61 is created by averaging pixels at the same position between these images or by cumulative addition so as to remove a change in density (hereinafter, referred to as a basic image).

Next, the basic image 61 and the TV camera 65
Then, a difference from the current image 62 input from the above is calculated, and a block 63 of pixels in the monitoring area is extracted. In this case, noise may be extracted.
When the area (the number of pixels) is equal to or larger than a predetermined threshold value, the object is extracted on the assumption that the monitoring target object has entered the monitoring area.

[0005]

The above-mentioned conventional apparatus has the following problems. (1) In outdoor monitoring, the gray value of an image input from a TV camera changes depending on the time, for example, morning, evening, or the weather, for example, fine, cloudy, or rainy. For this reason, when a basic image is obtained from an image accumulated only once for a certain period of time, the density value varies depending on the state at the time of accumulation, and an object that has entered the monitoring area cannot be accurately extracted.

(2) Objects other than the monitoring target object that has entered the monitoring area are sometimes extracted. (3) When there are a plurality of areas to be monitored, a shadow of an object that has entered an adjacent area may be extracted as an intruding object.

[0007] The present invention solves such a conventional problem, and in outdoor monitoring, it is possible to accurately monitor whatever the weather and time change, and to monitor in a monitoring area. It is an object of the present invention to prevent a malfunction from occurring even when a different object enters, and to prevent a malfunction due to a shadow of an object in an adjacent monitoring area.

[0008]

FIG. 1 is an explanatory view of the principle of the present invention, FIG. 1 (a) is an explanatory view of updating a basic image, and FIG. 1 (b) is an explanatory view of a configuration of a monitoring unit. It is.

In FIG. 1A, a monitoring area A and another area a around the area A are set in an image I from a monitoring TV camera. The area a is set smaller than the width of the monitoring area A, that is, the area is set smaller.

In the area a, the basic image is updated at regular time intervals Δt. For example, a basic image is created by averaging a predetermined number of images Ia ₁ , Ia _2, ... Ian in the area a, and this is stored in a memory or the like until the next basic image is updated.

In the explanatory diagram of the configuration of the monitoring unit shown in FIG. 1B, an image input unit 1 outputs image data S2 from a monitoring TV camera to a monitoring unit 4. The monitoring unit 4
An image data distribution unit 44 that distributes image data of the area a and the monitoring area A, an area monitoring unit 41a that monitors the area a, and a monitoring area monitoring unit 41b that monitors the monitoring area A are provided.

An area monitoring section 41a receives a processing start signal S1 at regular time intervals Δt by a timer or the like, and generates a basic image of an area a, a basic image generating section 411a, and an image of the basic image and an image of the current area a. Is provided with an object extraction unit 412a that detects an approaching object from the difference between the two.

The monitoring area monitoring section 41b includes an object extracting section 41
2a, a basic image generating unit 411b for generating a basic image of the monitoring area A, and an object extracting signal S3 by extracting an entering object from a difference between the basic image and the current image of the monitoring area A. Is provided.

The basic image generator 411b calculates the average of a predetermined number of images IA ₁ , IA ₂ ... IAn (not shown) in the monitoring area A in the same manner as the basic image generator 411a. By obtaining, the basic image is created.

The basic image generators 411a and 411a
The number of images predetermined for averaging or cumulative addition in 11b can be smaller for those with a fast approaching speed of the object, and can be increased for slower ones.

[0016]

The area monitoring section 41a of the monitoring section 4 receives the processing start signal S1 at regular time intervals Δt,
1a, a basic image in the area a is generated and the basic image is updated.

An intruding object is detected from the difference between the updated basic image and the image of the current area a by the object extracting unit 412a, and a detection signal is output. The monitoring area monitoring unit 41b
Upon receiving the detection signal of the entering object, the basic image generation unit 411
The base image of the area A is updated by b.

An intruding object is extracted from the difference between the updated basic image and the current image of the monitoring area A by the object extracting unit 412b, and an object extraction signal S3 is output. Therefore, since the area of the region a is small, the basic image of the region a can be updated at a high speed, and the update time interval can be shortened. In addition, when an intruding object is detected in the area a, the basic image of the monitoring area A is updated in response to the detection of the intruding object. It is possible to accurately monitor even if the weather and time change.

[0019]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to FIGS. (1) Description of the monitoring area in the first embodiment FIG. 2 is an explanatory diagram of the monitoring area, FIG. 2 (a) is an actual place in outdoor monitoring, and FIG. 2 (b) is a description of setting of the monitoring area. FIG.

In FIG. 2A, a plurality of elongated objects X
Are entered one by one from the direction in which the areas a1, a2, and a3 for object detection are provided by arrows, that is, a plurality of monitoring areas A
This is monitoring when stopping in a row alongside 1, A2, and A3. TV camera 5 for monitoring the place where this object X stops
Monitor when and where the object X has entered.

FIG. 2B is an explanatory diagram for setting a monitoring area on the image I from the TV camera 5. FIG. 2 (b)
, A plurality of monitoring areas, for example, three monitoring areas A
1, A2 and A3 are provided. And these monitoring areas A
Object detection areas a1, a2, and a3 are provided in directions in which the object X near 1, A2, and A3 enters, respectively. The areas of the areas a1, a2, and a3 are the monitoring areas A1, A2, and A3.
It is set smaller than. Further, each of the monitoring areas A1, A2, and A3 is divided into two monitoring areas A11.
And A12, A21 and A22, and A31 and A32.

In the image I, the surveillance area A1 is large and the surveillance area A3 is small.
This is because the closer to the camera the larger the size.

(2) Description of the Overall Configuration of the Monitoring Device in the First Embodiment FIG. 3 is an explanatory diagram of the overall configuration of the monitoring device. 3, the monitoring device includes an image input unit 1, a timer 2, a time management unit 3, three monitoring units 4 corresponding to the monitoring areas A1, A2, and A3 in FIG. 2B, and a TV camera 5. .

The image input unit 1 receives an analog image signal from the TV camera 5, performs A / D conversion, and digitizes the signal. The digitized digital image is stored in a frame memory (not shown), for example, 256 pixels × 256 pixels × 8
Each of the monitoring units 4-1 and 4-
2, 4-3.

The timer 2 has a clock function and not only outputs the current time to the time management unit 3 but also updates the basic images of the object detection areas a1, a2 and a3 (described later).
Therefore, each monitoring unit 4-
A signal S1, for example, a square pulse is sent to 1, 4-2, and 4-3.

The time management unit 3 includes a plurality of monitoring units 4-1 and 4-
Signals S3-1, S3-2, S3 sent from 2, 4-3
-3 indicates to the user which signal is sent from which monitoring unit and when. For example, a table is created in which three monitoring units 4-1 4-2, and 4-3 are associated with the monitoring areas A1, A2, and A3, and a lamp is turned on in a column corresponding to the monitoring unit that sent the signal. Then, the time sent from the timer 2 is displayed based on the sent time.

The monitoring units 4-1, 4-2, 4-3 are provided by the number corresponding to the monitoring areas A1, A2, A3, and in this embodiment, all three have the same configuration. Each monitoring unit 4
-1, 4-2, and 4-3 are processing start signals S1
And an input object X is extracted by inputting image data S2 from the image input unit 1 and object extraction signals S3-1, S3-2, S
Output 3-3.

The operation of this monitoring device is performed by monitoring units 4-1 and 4
The processing start signal S1 is input from the timer 2 at fixed time intervals set by the user in -2 and 4-3. Thus, the monitoring units 4-1 4-2, and 4-3 monitor the entering object X from the image data S2 from the image input unit 1. When the monitoring units 4-1, 4-2, and 4-3 extract the approaching object X, the monitoring unit 4 outputs the object extraction signals S 3-1, S 3-2, and S 3-3 to the time management unit 3. Thereby, three monitoring areas A1,
The user is notified when and where the object has entered A2 and A3.

(3) Description of the monitoring unit in the first embodiment FIG. 4 is an explanatory diagram of the monitoring units 4-1, 4-2, and 4-3.
Since these have the same configuration, the monitoring unit 4-1 will be described as a representative. FIG. 4A is an overall configuration diagram of the monitoring unit 4-1, and FIG. 4B is an explanatory diagram of the monitoring module 41a.

In FIG. 4A, a monitoring unit 4-1 for monitoring a monitoring area A1 includes three monitoring modules 41a, 4a, and 4b.
1b, 41c, two registers 42a, 42b, a comparison unit 43, and an image data distribution unit 44.

The monitoring module 41a uses the signal from the timer 2 as a processing start signal S1, and
4 from the image data of the area a distributed from
When an object X entering the vehicle is detected, a detection signal of the object is output to the monitoring module 41b.

The monitoring module 41b uses the detection signal of the monitoring module 41a as a processing start signal and extracts an object X entering the monitoring area A11 from the image data of the monitoring area A11 distributed from the image data distributor 44. If so, the object extraction signal is output to the monitoring module 41c and the register 42a.

The monitoring module 41c uses the object extraction signal of the monitoring module 41b as a processing start signal, and uses the image data of the monitoring area A12 distributed from the image data distribution unit 44 to determine the object X entering the monitoring area A12. If extracted, the object extraction signal is output to the register 42b.

The registers 42a and 42b hold the respective object extraction signals of the monitoring module 41b and the monitoring module 41c. The comparing unit 43 compares the two registers 42a and 42b, and outputs the monitored object extraction signal S3-1 when the two registers both hold the object extraction signal.

The image data distribution unit 44 converts the image data S2 from the image input unit 1 into the corresponding monitoring modules 41a, 41b, based on the setting of the area a1, the monitoring areas A11, A12 shown in FIG. The image data is distributed to 41c.

The operation of the monitoring section 4-1 is as follows. When the monitoring module 41a detects the entering object X, the monitoring module 41b receives the object detection signal and then performs the object monitoring operation. . The object extraction signal of the monitoring module 41b is stored in the register 42a. For example, if an object is extracted by the monitoring module 41b, the data of "1" is stored in the register 42a, and the data of "0" is stored in the other registers.

Further, after the monitoring module 41c receives the object extraction signal from the monitoring module 41b, the monitoring module 41c performs an object monitoring operation. The object extraction signal output from the monitoring module 41c is
b. This storing method is the same as that of the register 42a.

In the comparing section 43, the two registers 42
Referring to a and 42b, if the data is both "1", an object extraction signal S3-1 indicating that the monitoring target object has been extracted is output as the output of the monitoring unit 4-1.

Thus, when both the monitoring module 41b and the monitoring module 41c extract an object, the object extraction signal S3-1 is output, so that the approach of an elongated object such as an airplane can be monitored.

FIG. 4B is an explanatory diagram of the monitoring module 41a. The monitoring modules 41b and 41c have the same configuration as the monitoring module 41a, and the same module can be used.

In FIG. 4B, the monitoring module 4
1a denotes a basic image generation unit 411, an object extraction unit 412, and an RO
M413 and two frame memories 414 and 415.

FIG. 5 shows the basic image generator 4 shown in FIG.
FIG. 11 is a detailed explanatory diagram of an object extraction unit and an object extraction unit; In FIG. 5, a basic image generation unit 411 includes a frame memory 414
Clearing unit 4111, adding unit 4112, and control unit 4113
And a division unit 4114.

The object extracting unit 412 is provided in the frame memory 41
5, a clearing unit 4121, a subtracting unit 4122, a labeling unit 4123, and an area calculating unit 4124. The operation of the monitoring module 41a is shown in FIGS.
It will be described based on.

As shown in FIG. 5, the basic image generation unit 411 uses the processing start signal S1 from the timer 2 input to the monitoring module 41a as a start signal, and first, the clear unit 41
In step 11, the contents of the frame memory 414 external to the basic image generation unit 411 are all cleared to "0". The frame memory 414 has a depth of, for example, 16 bits for adding image data to a depth of 8 bits of the memory indicating the density, color tone, and the like of the frame memory of the image input unit 1.

After clearing the frame memory 414, the adder 4112 sequentially adds the input image data of the area a1 from the image data distributor 44 to the data in the frame memory 414, and adds the added value to the frame memory 414.
14 is held.

The control section 4113 repeats the addition processing in the addition section 4112 according to the addition number stored in the ROM 413 in the monitoring module 41a of FIG. 4B. After completion of the addition, the pixel values in the frame memory 414 are each divided by the number of additions by the division unit 4114 for averaging, and the average value is used as a basic image in the frame memory 4.
14 is held. Specifically, the division unit can be realized by a shift register by setting the number of times of repeating addition to a power of two.

After completion of the division, the basic image generation unit 411 outputs a start signal to the object extraction unit 412. Object extraction unit 4
In step 12, the contents of the frame memory 415 externally attached to the object extraction unit 412 by the clear unit 4121 are first cleared to “0” using the output from the basic image generation unit 411 as a processing start signal.

Next, the subtractor 4122 subtracts the contents of the frame memory 414 from the input image from the image data distributor 44 for each pixel. If the difference is negative, “0” is stored in the frame memory 415, and the shadow portion of the object is ignored.
This is because the object is assumed to be brighter than the road surface, and if the object is negative, it is determined to be a shadow. When the difference is equal to or less than a certain value, “0” is similarly stored. When the difference is equal to or more than a certain value, “1” is recorded in the frame memory 415.

In the portion of "1" recorded in the frame memory 415, a continuous area is defined by the labeling section 41.
23, and assigns the same label number to the obtained continuous area.

The area calculation unit 4124 counts the number of pixels for each label number of the above-described area, and if any one area exceeds a certain number, determines that an object has been extracted and outputs an object extraction signal. Output.

Thus, an object detection signal is output from the monitoring module 41a to the monitoring module 41b. In the above embodiment, the monitoring areas A1, A
2. Although A3 is limited to three, a large number of areas can be monitored by increasing the number of monitoring areas and monitoring units 4. Also, in the subtraction unit 4122, contrary to the above embodiment, the input image from the image data distribution unit 44 can be subtracted for each pixel from the contents of the frame memory 414. In this case, if the difference is positive or equal to or greater than a certain value, “0” is recorded in the frame memory 415 if it is equal to or less than a certain value.

(4) Description of the Monitoring Area Considering the Shape of the Monitoring Object in the Second Embodiment FIG. 6 is an explanatory diagram of the monitoring area considering the shape of the monitoring object, and FIG. FIG. 6B is an explanatory diagram for monitoring an object, and FIG. 6B is an explanatory diagram for monitoring a rectangular object.

In FIG. 6A, the monitoring area A1 is divided into two monitoring areas A11 and A12 and arranged side by side, and object detection areas a1-1 and a1-2 are provided on the left and right of the monitoring area A1.
Is provided.

The areas a1-1 and a1-2 can be any one if the approach direction of the object is determined. For example, if the approach direction of the object is the direction of the area a1-1, when the object is detected in the area a1-1, the monitoring area A1
1 starts monitoring. When an object is extracted in the monitoring area A11, monitoring of the monitoring area A12 is started. And
When an object can be extracted in both the monitoring area A11 and the monitoring area A12, it is possible to extract the entry of an elongated monitoring target object such as an airplane, for example, for the first time.

When the approach direction of the object is determined as described above, the monitoring area can be limited to A11 at first, but when the approach direction of the object is not determined, the area a1-
When the object is detected in 1 or a1-2, the monitoring area A1 is always
1 and A12 are simultaneously monitored, or from the monitoring area A11 to A12 or the monitoring area A
Monitoring is performed alternately from 12 to A11.

FIG. 6B is an explanatory diagram for monitoring a rectangular object. The monitoring area A1 is divided into four monitoring areas A11,
A12, A13, and A14 are arranged separately, and object detection areas a11-1, a11-2, and a1 are provided around the monitoring area.
2-1, a12-2, a13-1, a13-2, a14
-1, a14-2.

The four monitoring areas A11, A12, A1
3. When the objects are extracted from all the monitoring modules provided corresponding to A14, it is possible to extract the entry of a rectangular object.

The monitoring areas A11, A12,
The number and arrangement shape can be changed according to the shape of the monitored object. (5) Description of Detection of Passage of Object in Third Embodiment FIG. 7 is an explanatory diagram of detection of the passage of an object, and FIG. 7 (a) monitors the passage of a horizontally long object X1. FIG. 7B is an explanatory diagram for monitoring the passage of a vertically long object X2.

In FIG. 7A, a plurality of monitoring areas A
1, A2, A3 and areas a1, a2, a3 are arranged side by side in the approach direction of the object X1, for example, as illustrated. The monitoring units corresponding to the monitoring areas A1, A2, and A3 use the object X
When 1 is extracted, it can be detected that the object X1 has passed through the monitoring areas A1, A2, and A3, which are predetermined passing areas, in the direction of the arrow.

FIG. 7B is an explanatory diagram for monitoring the passage of a vertically long object X2. In FIG. 7B, a plurality of vertically long monitoring areas A1, A2, A3 and areas a1, a2, a
3 are arranged side by side in the approach direction of the object X2, for example, as shown in the figure.

In this case as well, as in FIG.
When the object X2 is extracted by the monitoring units corresponding to 1, A2, and A3, the object X2 is moved from the direction of the arrow to the monitoring area A1,
It is possible to detect the passage of A2 and A3.

In the above embodiment, the basic images such as the areas a1 and a2 for detecting the entry of an object are always updated at regular time intervals. If so, the update of the basic images such as the areas a1 and a2 can be started at that time.

[0063]

As described above, the present invention has the following effects. (1) An area provided around the monitoring area and having a small area for detecting entry of an object is provided, and a basic image of the monitoring area is updated by a detection signal of an object in this area, and monitoring is performed based on the basic image. Therefore, the outdoor monitoring can always perform accurate monitoring regardless of the weather or time.

(2) By providing the monitoring area divided into a plurality of areas, it is possible to prevent malfunction due to entry of an object different from the monitoring target. (3) Since a difference between the basic image and the current image is obtained and a portion where the value of the current image is smaller than the value of the basic image is not processed, a malfunction due to a shadow of an object in an adjacent monitoring area can be prevented.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is an explanatory diagram of a monitoring area according to the first embodiment of the present invention.

FIG. 3 is a configuration diagram of an apparatus according to a first embodiment of the present invention.

FIG. 4 is an explanatory diagram of a monitoring unit according to the first embodiment of the present invention.

FIG. 5 is an explanatory diagram of a basic image generation unit and an object extraction unit according to the first embodiment of the present invention.

FIG. 6 is an explanatory diagram of a second embodiment of the present invention.

FIG. 7 is an explanatory diagram of a third embodiment of the present invention.

FIG. 8 is an explanatory diagram of a conventional example.

[Explanation of symbols]

a region A monitoring region I image Ia ₁ , Ia ₂ , Ian region a image 1 image input unit 4 monitoring unit 41a region monitoring unit 411a basic image generation unit 412a object extraction unit 41b monitoring region monitoring unit 411b basic image generation unit 412b object Extractor

────────────────────────────────────────────────── ─── Continuation of front page (56) References JP-A-5-205054 (JP, A) JP-A-5-314392 (JP, A) JP-A-3-144797 (JP, A) (58) Field (Int.Cl. ⁷ , DB name) G08G 1/04 G06T 1/00 G08B 13/194 H04N 7/18

Claims (6)

(57) [Claims] 1. An object monitoring apparatus comprising: an object extracting unit that compares a basic image generated by a basic image generating unit with a current image to extract an object that has entered a monitoring area (A); An area monitoring unit (41a) for detecting an object that has entered the area (a) around the area (A) is provided, and the basic image generation unit (411b) of the monitoring area (A) is provided with a signal from the area monitoring unit (41a). An object monitoring device, wherein a basic image of the monitoring area (A) is updated by an object detection signal. 2. A basic image generating section (411a) for generating a basic image of the peripheral area (a) at regular time intervals in the area monitoring section (41a), and a basic image of the peripheral area (a). provided object extraction unit for detecting an object that enters into the area of the peripheral by comparing the current image (a) of generating area of the peripheral and generated basic image (411a) (a) and (412a) The object monitoring device according to claim 1, wherein: 3. The object monitoring apparatus according to claim 1, wherein the area of the peripheral area (a) for detecting the entry of the object is set smaller than the area of the monitoring area (A). 4. The difference between the basic image of the monitoring area (A) and the current image is obtained, and the brightness of the current image is calculated based on the basic image.
The monitoring that does not process a portion that is less than a certain value from the brightness
2. The object monitoring apparatus according to claim 1, further comprising an object extraction unit (412b) for the area (A) . 5. The object monitoring apparatus according to claim 1, wherein the monitoring area is divided into a plurality of areas according to the shape of the monitored object. 6. The object monitoring apparatus according to claim 4, wherein a plurality of said peripheral areas (a) are arranged in accordance with the approach direction of the object.