JP2003299074A - Variation detecting apparatus and detecting method using camera apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve the problem that a monitoring apparatus responds sensitively to a small motion such as window vibration or material vibration by wind, light incidence condition, a change of brightness in morning, noon, and night, a change of light quantity by turning ON/OFF a lamp, the existence of a moving material such as a metronome on the screen, and it may cause a faulty detection. <P>SOLUTION: A timer or a counter for monitoring motion condition is provided at the monitoring apparatus. The counter performs an update operation of a reference screen for an unmoved screen, and operates to reset the counter sequentially for a screen where a motion is detected, based on the result of motion determination. According to this arrangement, the value of the counter does not reach a maximum value unless an unmoved condition continues for a predetermined period of time or more. Therefore, a screen where the maximum value is reached can be graded as the unmoved screen, and this screen is employed as a new reference screen to update an old reference screen. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a change detecting device and a detecting method using an image pickup device, and more particularly to a monitoring device and a monitoring method such as a monitoring device capable of accurately detecting a change in movement.

[0002]

2. Description of the Related Art A monitoring device using an image pickup device is, for example,
Many are used for the purpose of monitoring intruders or intruding objects (hereinafter referred to as intruding objects) inside stores such as banks, supermarkets and convenience stores, construction sites, and restricted areas such as dams, bases and airfields. ing.

As a method of detecting an intruding object in such a monitoring device, for example, a window area for motion detection is provided in a screen imaged by an imaging device, and video information in the area is periodically detected. There is a device which, when compared with a reference signal of a screen photographed in the past and detects a difference of a certain amount or more, judges that there is motion and sounds an alarm.

As another method, an image of a place to be monitored is captured by an image pickup device in advance, this is recorded as a reference screen, and this screen is compared with the screen captured at the time of monitoring to determine whether or not there is any movement. There is a motion detection device for judging.

Further, as a well-known system, a method of detecting a motion by detecting a difference between frames or fields of an image on an image pickup screen is well known. FIG. 4 shows a well-known monitoring device.

In FIG. 4, 41 is an image pickup device, and 42 is an image pickup device.
Is a control unit, 43 is an image processing unit, 44 is an image storage unit, and 45 is a monitor. The operation of this device will be described with reference to the flowchart shown in FIG. First, the image captured by the image capturing device 41 is the video signal input I
Is input to the control unit 42 and recorded in the image storage unit 44. In the image processing unit 43, the previously stored screen, for example, the screen one frame or one field before is registered as the reference screen 51. In this registered image, the brightness of the background changes with the passage of time, so compare the brightness between frames or fields as appropriate,
It is also well known that the reference screen is automatically updated when the brightness difference exceeds a predetermined brightness difference.

Next, a case where an intruder invades a predetermined monitoring place will be described. The current screen 52 in which an intruder has entered the field of view of the image pickup device is used as the image signal input I by the image processing unit 43.
Entered in. In the image processing unit 43, the reference screen 51 and the current screen 52 stored in advance are 53, and the block unit difference calculation processing is performed. In this block unit difference calculation processing 53, for example, the screen is divided into a plurality of blocks, and for example, the average value of the brightness values is compared between each reference screen 51 and the current screen 52 for each block, and the brightness difference value is calculated. Calculated. Here, one block is a unit of picture element and is 4 ×.
4, 8 × 8, 16 × 16, ... Are often used.

The calculation result calculated in the block unit difference calculation processing 53 is integrated in 54 and grasped as the amount of movement of the intruder. The output of the integration process 54 is motion-determined at 55. The motion determination 55 binarizes the output of the integration processing 54 and compares it with a predetermined threshold value. If the motion amount is larger than a predetermined threshold value, it is determined that there is an intruder,
It is displayed on the monitor 45 as an alarm.

In fact, there are many cases where there is a small movement at the place to be monitored even if there is no intruder. For example, if a surveillance camera is installed at the entrance of a supermarket or convenience store,
Alternatively, the surrounding windows are often transparent, and even when there are no customers in the store, outside movement is often in the field of view of the surveillance camera through the door or through the window, and with simple difference detection, There is an erroneous detection or erroneous determination that there is motion.

Also, small movements such as window vibrations and object vibrations due to the wind, how the light shines, changes in brightness in the morning, noon, and evening, and changes in the amount of light due to ON / OFF of the light,
Even when there is a constantly moving object such as a metronome in the screen, it reacts sensitively, and these often cause erroneous detection.

[0011]

As described above,
In the case of a method that takes the difference between the current screen and the previous screen as the reference image, even if there is no intruder, it reacts sensitively to even slight changes in the object within the field of view of the image pickup device, and the movement is considered as movement. In many cases, it makes an erroneous detection of detecting and issuing an alarm. Further, in the case of this conventional method, the movement cannot be detected at the moment when the movement is stopped, for example, the intruder stops. However, even in such a case, the monitoring device needs to issue an alarm indicating that there is an intruder and record the alarm. As described above, in the conventional monitoring device, when an intruder is present and is significantly different from the reference image, the alarm is continuously issued. However, for example, even when there is a change in the light amount due to ON / OFF of the light, it is different from the reference image. However, even if there is no intruder, an alarm will continue to be issued, and if there is an intruder and the user stops and does not move, there is a problem that an alarm cannot be issued and correct detection cannot be performed.

An object of the present invention is to provide a monitoring device with extremely high accuracy.

Another object of the present invention is to provide a monitoring device in which malfunctions due to minute fluctuations or sudden fluctuations are minimized.

Still another object of the present invention is to provide a monitoring device for detecting an intruding object even if the intruding object stops moving.

[0015]

In order to achieve the above object, the present invention has a function of constantly updating the reference screen to an optimum one. In order to realize this, a screen difference is always taken for each screen and the result is recorded. Based on the result, if there is intruder,
Depending on the movement, you can judge whether there is movement or not,
Since each screen changes rapidly, a timer or counter is provided to monitor the state of motion. This counter operates based on the motion determination result. That is, when the count value of the counter reaches the maximum value M,
Operates to update the reference screen. Therefore, the reference screen is updated when there is no motion, and the counter is sequentially reset when the motion is detected. That is, since the count value does not become M, the reference screen is not updated. In this way, the value of the counter does not reach the maximum value M unless the state of no motion continues for a certain period. Therefore, a screen whose count value has reached the maximum value can be regarded as a screen without motion. Therefore, adopt this screen as the reference screen,
Update the old standard screen. In this way, when an intruder exists, it is very rare that the intruder stays in the same state for a long time. That is, in a moving screen, the counter is reset before the maximum value M is reached, and the reference screen is not updated while the screen without an intruder is stored.

The reference screen is determined by such a method, and a real-time difference between the reference screen and the reference screen is calculated. In this way, when there is an intruder, the reference screen is the image before the intrusion, and therefore the screen on which the intruder is present shows a large difference value even if the intruder has little movement. As a result, it is possible to keep issuing the alarm. Also, when the light is turned on (or off) and the light amount changes significantly, the difference value shows a large value at that moment, but after that, the reference screen switches to the image of the light on (or off), so that the correct reference is immediately obtained. It becomes a screen.

By doing so, since the motion detection is always carried out based on the optimum reference screen, it is possible to prevent erroneous detection of the monitored object or image pickup error such as missing of the image to be detected.

[0018]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described in detail below with reference to FIGS. 1 is a diagram showing a schematic configuration of a monitoring device of the present invention, FIG. 2 is a flow chart diagram for explaining the operation of the monitoring device shown in FIG. 1, FIG. 3 is a signal waveform, and the horizontal axis is time. Show progress.

In FIG. 1, reference numeral 101 denotes a surveillance TV camera, which is installed in various places such as a store and a restricted area. Although one unit is shown in the figure, a plurality of units are generally installed. A monitor 102 displays a video image or a recorded image from the surveillance TV camera 101. 103
Is a control device, and is composed of blocks 104 to 113 and a bus 110 connecting these blocks. 104
Is a communication control unit that controls the surveillance TV camera 101,
A video interface 105 has an image input interface (not shown) and an image signal output interface (not shown). Reference numeral 106 denotes a recording device for recording an image from the surveillance TV camera 101, 107
Is a frame buffer memory, 109 is an input device for various settings, 112 is a memory, 113 is an image processing unit, and 111 is a signal processor (CPU) that controls the components of each of these blocks.

Video signals taken by the surveillance television camera 101 are recorded in the storage device 106. The recorded image is processed by the image processing unit 113 to create a reference image, which will be described in detail with reference to FIG. In the image processing unit 113, the data of the previous screen, for example, the data 201 divided into a plurality of blocks in the image data of one field before is read from the storage device 106. Here, a plurality of blocks means, for example, 4 (pixels) × 4 (pixels)
It is possible to divide a block (hereinafter abbreviated as 4 × 4) into one block, 8 × 8, 16 × 16, or the like.

On the other hand, the current image picked up by the surveillance television camera 101 at present is temporarily stored in the frame buffer 107, and the data 202 divided into a plurality of blocks is read out from the image data of one field. Here, the data 202 divided into a plurality of blocks is the data 201 divided into a plurality of blocks of the previous image data.
It is a block with the same size and position as.

The block data is detected as various data for each block, such as an average value of luminance, an average value of hue, an average value of saturation, an integrated value of high frequency components, etc. In this embodiment, in order to facilitate understanding, First, the processing with the average luminance value will be described below. Block division data 201 of previous screen
And a luminance average value of each of the block division data 202 of the current screen is calculated by a block unit difference calculation 203 between the luminances of both screen data. This calculation is performed in all blocks of one field or all blocks of one frame.

At this stage, even a slight movement of the object on the screen is detected as the amount of movement, so there is a high possibility that noise will be unnecessary for detection. In order to remove small movements such as noise in this way, two-dimensional filtering 2
Filtering processing is performed at 04 to remove unnecessary noise. Integration processing 205 is performed on the data after noise removal. This integral value is the integral value of the difference between the brightness values of the previous screen and the current screen as described above, and reflects the magnitude of the motion. FIG. 3 shows such a situation. In the state (A), if there is no intruder during the T ₁ period and there is an intruder during the T ₂ period, the output signal of the integration processing 205 is the motion amount D. Can be expressed as

The output signal of the integration processing 205 is filtered by a time axis filter 206 known in the related art in order to remove a sudden change in motion. next,
The result of the time-axis filtering process is binarized, and the motion determination 207 determines the motion. In this motion determination, a predetermined reference value (threshold h ₀ ) is set, the motion determination is performed based on whether the value is larger or smaller than the reference value, and a binary result is output. This threshold value h ₀ may be set by making an experimental determination in advance and set by the operator or while looking at the monitor 102. As a result, the motion determination signal J is output and sent to the status monitor 208.

The operation of the status monitor 208 will be described. The state monitor 208 has a state monitor counter 108 (shown in FIG. 1), and the state monitor counter 108 is preset with a maximum value M. As shown by the output of the state monitoring counter C in FIG. 3, when the output result J of the motion determination 207 is a motionless determination, the output result J is counted up for each screen (or for each field) and the output result J In that case, reset processing is performed. Here, the counter can reach the maximum value M only when no motion continues for N consecutive screens. Here, the values of N and M can be set experimentally depending on the situation of the monitoring place.

When the state without motion continues for N consecutive screens, the counter reaches the maximum value M. In this case, since it can be regarded as a steady state, the screen at that time is adopted as the reference screen. That is, looking at the update status of the reference screen S ₀ , when the state monitoring counter 108 reaches the maximum value, it outputs the reference screen update instruction SS. As a result, the image a that was the reference screen until immediately before that (held in the memory 112)
Rewrites the memory 112 as the image b. Similarly, when there is no motion for N consecutive screens,
The image b is rewritten to the image c. In this way, the reference screen is updated.

Next, the method of motion detection using this reference screen will be described below. Of the image data of one field of the updated reference screen, the data 209 divided into a plurality of blocks is read from the memory 112. On the other hand, among the current images taken by the TV camera 101, as described above,
For example, of the image data of one field of the current screen, the data 202 divided into a plurality of blocks is read from the frame buffer 107, and the difference between the average luminance values of the block data of the current screen and the block data of the reference screen is in block units. It is calculated by the difference calculation 210. Although the average luminance value is used here, it goes without saying that data such as an average hue value, an average saturation value, and a high-frequency component integral value can be used as described above.

The result calculated by the block unit difference calculation 210 is filtered by the two-dimensional filter processing 211, and small movements such as noise are removed. When the data after noise removal is integrated by the integration processing 212,
It becomes a signal of the movement amount AD (shown in FIG. 3). This signal is a signal that reflects the magnitude of movement. The signal of motion amount AD is
Similarly to the above, in order to remove a sudden change in motion, the time axis filter 213 performs filtering processing. The result is binarized, and motion determination is performed in motion determination 214. In the motion judgment, a predetermined judgment standard (threshold h ₁ ) is set, and the motion is judged depending on whether it is larger or smaller than the judgment standard value, and an alarm is displayed on the monitor 102, for example. The monitor 102 displays the image from the TV camera 101, but if necessary, the image from the storage device 106 or an alarm can be superimposed and displayed.

The above operation will be described again with reference to the reference image update and motion detection time chart of FIG. In state A, when switching from state T ₁ without an intruder to state T ₂ with an intruder, the difference value (movement amount) of the brightness of the screen before and after switching is a large value when the intruder enters the screen. Indicates. However, since the comparison is between the current screen and the previous screen, the brightness difference value (movement amount) becomes smaller when the intruder stops moving. Figure 3
This is the part that has been shown to stop moving. In the motion determination, when the brightness difference value (motion amount) is equal to or greater than the threshold h ₀ , it is determined that there is motion, and thus there is a case where there is no motion in a portion lower than the threshold h ₀ , a motion stop portion in FIG. Indicates that.

The state monitor counter 108 repeatedly reaches the maximum value M in N screen cycles in a state where there is no movement, as in the output signal C of the state monitor counter. When the maximum value M is reached, the reference screen is updated. Image b and image c correspond to this. After the image c is updated, a movement occurs when an intruder enters, and the state monitoring counter 108 is reset before the maximum value M is reached. When the movement of the intruder stops, that is, when the movement amount D becomes lower than the threshold value h ₀ , the state monitoring counter 108 counts up but does not continue for a long time. Therefore, if N is set to a predetermined value, the state monitoring counter 108 Is reset before reaching the maximum value M, and eventually,
Until there is no intruder, the reference screen, that is, the image c
Is not updated.

On the other hand, the actual motion detection is performed by the difference between the reference screen and the current screen. Therefore, the difference in the brightness value between the screen in which the intruder entered and the screen before the intruder entered, that is, the image c in FIG. 3, is always large. Therefore, even if the intruder stops moving, the amount of difference in luminance value is large. Therefore, the determination of the motion is continued while the intruder is present in the screen, and the alarm is output.

The present invention has been described in detail above.
The present invention is not limited to the surveillance camera and the embodiment of the motion detection described here, and can be widely applied to an apparatus and a detection method for detecting a change in the field of view of the image pickup apparatus in addition to the above. It goes without saying that you can do it.

[0033]

According to the present invention, the reference screen is updated at a predetermined cycle when there is no motion, and the reference screen is not updated when there is motion. Therefore, motion detection is always performed based on the optimum reference image. Therefore, it is possible to realize an extremely practical monitoring device without erroneous detection or missing of an image to be detected.

[Brief description of drawings]

FIG. 1 is a block diagram for explaining an embodiment of the present invention.

FIG. 2 shows a flow chart for explaining the operation of the present invention.

FIG. 3 shows a time chart for explaining the operation of the present invention.

FIG. 4 is a diagram showing an example of a conventional monitoring device.

FIG. 5 shows a flowchart for explaining the operation of a conventional monitoring device.

[Explanation of symbols]

101: TV camera, 102: Monitor, 103: Control device, 104: Communication control unit, 105: Video interface, 106: Storage device, 107: Frame buffer,
108: status monitoring counter, 109: input device, 11
0: bus, 111: CPU, 112: memory, 113: image processing unit

─────────────────────────────────────────────────── ─── Continuation of front page (51) Int.Cl. ⁷ Identification code FI theme code (reference) H04N 5/225 H04N 5/225 C F term (reference) 5B057 AA19 BA02 CE06 DA15 DB02 DB09 DC32 5C022 AA01 AC69 5C054 AA01 CG06 CH02 EA01 EA05 ED07 EF06 FC01 FC05 FC13 GB11 HA18 5C087 AA02 AA03 AA04 AA09 AA24 DD05 DD20 DD49 EE14 GG06 GG19 GG37 5L096 AA06 BA02 CA04 DA03 GA08 GA09 GA19 GA51 GA55 HA01

Claims (7)

[Claims] 1. An imaging device for imaging an object, a recording device for recording an output video signal of the imaging device, an image processing unit for processing the output video signal, and a processing result of the image processing unit. An output device and the image pickup device, the recording device, the image processing unit, and a control device that controls the output device. The image processing unit is obtained from the previous screen obtained from the image pickup device and the image pickup device. The control device updates the reference screen when the output level of the comparison result from the image processing unit is below the first predetermined level for a predetermined period. If the output level of the comparison result from the image processing unit is equal to or higher than the first predetermined level within a predetermined period, the reference screen is not updated. 2. The change detection device according to claim 1, wherein the control device has a state monitoring counter, and outputs a command for updating the reference screen when the state monitoring counter reaches a predetermined value. Change detection device. 3. The change detection device according to claim 1, wherein the image processing unit further has a function of comparing the updated reference screen with a current screen obtained from the imaging device, The change detection device, wherein the control device has a function of determining that there is a change when the output level of the comparison result from the image processing unit exceeds a second predetermined level. 4. The change detection device according to claim 2, wherein the image processing section performs a difference calculation between a previous screen obtained from the image pickup device and a current screen obtained from the image pickup device, The calculation result is subjected to a two-dimensional filter processing, and the result of the above two-dimensional filter processing is subjected to an integration processing to calculate a motion amount,
A change detection apparatus having a function of performing a time axis filter process on the above-described motion amount calculation result to determine a motion. 5. The change detection device according to claim 3, wherein the image processing unit further calculates a difference between the updated reference screen and the current screen obtained from the image pickup device, and outputs the calculation result. Is subjected to a two-dimensional filter process, the result of the two-dimensional filter process is integrated, a motion amount is calculated, and the motion amount calculation result is subjected to a time axis filter process. A change detection device characterized by generating an alarm signal based on a result. 6. A step of imaging an object with an imaging device,
The step of recording the output video signal of the image pickup device, the step of comparing the previous screen obtained from the image pickup device and the current screen obtained from the image pickup device and the motion amount of the comparison processing result are It consists of a judgment step,
The step of determining the amount of movement includes the step of updating the reference screen if the amount of movement is below a predetermined level for a predetermined period, and the amount of movement of the comparison processing result within a predetermined period,
A change detecting method comprising a step of not updating the reference screen when the level is equal to or higher than a predetermined level. 7. The change detecting method according to claim 4, further comprising the step of comparing the updated reference screen with a current screen obtained from the image pickup device, and the output level of the comparison result is A change detecting method comprising a step of determining that there is a change when a predetermined level of 2 is exceeded.