JPH0520563A - Smoke detector using image processing - Google Patents

Abstract

PURPOSE:To perform the fire judgment to make compatible the early discovery and the erroneous information prevention of the fire by grasping the smoking quantity by processing the image of the monitoring area of the smoke photographed by a high sensitivity camera. CONSTITUTION:A light emitting device 14 to synchronize the photographing action of a camera device 10 and to be provided at the facing position is light- emitted and driven, the photographed image signal is converted to a multi- gradation picture element signal, stored into frame memories 20A and 20B, the smoking area is detected from the multi-gradation picture element signal of one screen stored in the frame memories 20A and 20B, and by operating and integrating the smoke concentration for each picture element based on the multi-gradation picture element signal in the smoking area, the smoking quantity is calculated, and the fire is judged from the smoking quantity. From the change quantity of the previous smoking quantity and the present smoking quantity, the fire may be judged.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a smoke detection device using image processing for processing a smoke image captured by a high-sensitivity camera to judge a fire.

[0002]

2. Description of the Related Art In a smoke detector used in a conventional fire alarm monitoring device, a large number of smoke detectors are dispersedly installed on a ceiling surface of a monitoring area, and smoke flowing into a smoke detecting chamber inside the detector is installed. The scattered light of the light from the light emitting element is received by the light receiving element, and when a light receiving output exceeding a predetermined threshold value is obtained, it is judged as a fire and a fire detection signal is sent to the receiver.

[0003]

However, in the conventional smoke detector, the state of smoke in only a very small part dispersed in the warning area is spot-detected, and the smoke detector is not detected. As long as there is no fire immediately below, it will take time for the rising smoke to spread along the ceiling surface and then flow into the detector to exceed the specified concentration, and it will be affected by the shape and size of the monitored area. There is a problem that it is easy.

The problem that it takes a long time to detect a fire can be increased by lowering the threshold value for fire judgment, but if the sensitivity is increased, on the contrary, an erroneous operation is likely to occur due to cigarette smoke other than a fire or water vapor due to cooking. The present invention has been made in view of such a conventional problem, and by processing an image of a surveillance area captured by a camera, the smoke amount itself can be grasped to realize early detection of a fire and prevention of false alarm. An object of the present invention is to provide a smoke detection device by simple image processing.

[0005]

To achieve this object, the present invention is constructed as follows. The reference numerals in the drawings of the embodiments are also shown. The smoke detection device using the image processing of the present invention includes a camera device 10 that images a monitoring area and outputs a video signal, and a light emitting device 14 that is installed on the side facing the camera device 10 across a monitoring space 50. A synchronizing device 16 for driving the light emitting device 14 to emit light in synchronization with the photographing operation of the camera device 10, and a frame for converting image signals of at least two screens photographed by the camera device 10 into multi-gradation pixel signals and storing them. Memory 20A, 20B and frame memory 20A,
A smoke area detection unit 22 for detecting a smoke area from a multi-tone pixel signal for one screen stored in 20B, and each pixel based on the multi-tone pixel signal in the smoke area detected by the smoke area detector 22 The smoke density calculation unit 24 which calculates the smoke density for each and stores it in the smoke area frame memories 26A and 26B, and the smoke density signal for each pixel stored in the smoke area frame memories 26A and 26B are integrated to calculate the smoke amount. Smoke amount integration unit 2
8 and a fire determination unit 32 that determines a fire based on the smoke amount calculated by the smoke amount integration unit 28.

Further, instead of the smoke generation area detection section 22 which determines the pixel area for which the smoke density calculation processing is performed from the multi-gradation pixel signal, an area mask setting section 42 which sets a mask 60 which determines a predetermined area is provided. You may make it calculate the smoke density for every pixel based on the multi-tone pixel signal contained in the area | region of the mask set by the mask setting part 42, and may store it in the smoke area frame memories 26A and 26B.

Further, the smoke area frame memories 26A, 26
Provided is a change amount integration unit 30 that calculates a time-based smoke change amount by calculating the difference between the current smoke density signal stored in B and the previous smoke density signal for each pixel in the subtraction unit 38 and integrating the pixel.
The fire determination unit 32 may determine the fire based on the smoke amount calculated by the smoke amount integration unit 28 and the temporal smoke change amount calculated by the change amount integration unit 30.

Further, a distance detecting section 40 for detecting the distance to the smoke emitting area photographed by the camera device 10 is provided, and the smoke density calculating section 24 calculates the smoke density after correcting the size of the smoke image according to the distance. Therefore, the smoke amount can be detected more accurately.

[0009]

According to the smoke detecting apparatus using the image processing of the present invention having such a configuration, the surveillance space is photographed by the high-sensitivity camera device and is provided at a position facing the camera in synchronization with the photographing operation. The light emitting device is driven, and the smoke emitted by the light emitting device to the surveillance space is photographed with the backlight illuminated.

In the image processing for photographing, it is first checked whether or not there is a smoke emitting area in which smoke is projected on the screen. To check whether or not this smoke generation area exists, the multi-gradation pixel signal of the previous screen is compared with the multi-gradation pixel signal of the current screen, and the decrease of the multi-gradation pixel signal of this time is equal to or higher than a predetermined level and When the range exceeds a certain area, specifically, a certain number of pixels, it is determined that there is a smoking area, and the amount of smoking is determined by using this as a smoking area candidate to determine whether or not there is a fire.

Since the value corresponding to the amount of smoke generated in the surveillance space can be detected by the image processing as described above, the amount of smoke generated can be detected quickly and accurately as compared with the conventional fire determination based on the spot-like smoke density by the smoke detector. Can judge a fire. In addition, cigarette smoke other than fire, smoke from cooking, and water vapor can be clearly distinguished from the fact that the amount of smoke itself is smaller than that of fire, and more accurate by observing the amount of change in smoke over time. Can judge.

[0012]

1 is a block diagram of an embodiment showing a first embodiment of the present invention. In FIG. 1, reference numeral 10 is a high-sensitivity camera, and for example, a CCD color camera or a color video camera can be used. In addition, high-sensitivity camera 10
Uses an autofocus unit 12 that can automatically control the focus of a subject. The high-sensitivity camera 10 is installed at a position overlooking the surveillance area such as a room, and is installed, for example, on the upper part of the wall surface of the room.

At the position facing the high-sensitivity camera 10, L
A light emitting device 14 including a light emitting element such as an ED is installed. The high-sensitivity camera 10 and the light-emitting device 14 are the synchronization device 16
Thus, the light emission operation is performed in synchronization with the shooting operation at every predetermined image sampling cycle. Therefore, when the high-sensitivity camera 10 captures an image of the surveillance space 50, if the smoke 100 rises in the surveillance space 50, the smoke 100 emits the backlight illumination by the light emission of the light-emitting device 14 synchronized with the photographing operation by the synchronization device 16. High-sensitivity camera 10 when received
Takes an image of smoke 100.

The image signal photographed by the high-sensitivity camera 10 is, for example, an NTSC image signal, of which the luminance signal component is converted by the AD converter 18 into a multi-tone digital signal at a predetermined sampling period, The first frame memory 20A and the second frame memory 20B are alternately stored. That is, the image data for two times consecutive in time is stored in the first frame memory 20A and the second frame memory 20B as a multi-gradation pixel signal for each pixel.

A smoke emitting area detector 22 is provided following the first frame memory 20A and the second frame memory 20B. The smoke area detection unit 22 detects a smoke area from the screen information currently being processed. The smoke area is detected by comparing the multi-gradation pixel signals for each pixel of the current screen and the entire screen stored in the first and second frame memories 20A and 20B, and the multi-gradation pixel signal of this time is compared with the previous time. When the pixel portions that have decreased by a predetermined level or more and the pixel portions that have decreased by the predetermined level or more are concentrated and have a predetermined number of pixels or more, a region that satisfies this condition is detected as a smoking region.

The smoke area detection information detected by the smoke area detecting section 22 is provided to the smoke density calculating section 24. The smoke density calculation unit 24 extracts the multi-gradation pixel signal included in the smoke area detected from the screen currently stored in the first frame memory 20A, for example, and is a smoke density signal for each pixel. Convert to. FIG. 2 is a characteristic diagram showing the relationship between the smoke density and the multi-tone levels used in the calculation of the smoke density calculating section 24.

In FIG. 2, the multi-gradation level shown on the horizontal axis is expressed by normalizing the multi-gradation level when smoke density is 0 [% / m] to 1.0, which is effective in reducing the multi-gradation level. As a result, the smoke density increases. Referring again to FIG. 1, the density signal for each multi-gradation pixel signal calculated by the smoke density calculation unit 24 is stored in the first smoke area frame memory 26A or the second smoke area frame memory 26B. In this embodiment, since the density signal for the previous screen and the density signal for the current screen are required, two smoke area frame memories 26 are provided.
A and 26B are provided.

First and second smoke emitting area frame memory 2
A smoke amount integration unit 28 is provided following 6A and 26B. The smoke generation amount integrator 28 reads out, for each pixel, the density signal included in the smoke area of the current screen stored in one of the first and second smoke area frame memories 26A and 26B, and integrates it to read the smoke area. Calculate the amount. on the other hand,
Change amount integration unit 3 that calculates the change amount of the previous and present smoke amounts
0 is provided, and the change amount integration unit 30 uses the first and second smoke area frame memories 26A, 2A obtained by the subtractor 38.
The temporal change amount of the smoke generation amount is calculated by obtaining a signal obtained by subtracting the density signal of the previous screen from the density signal of the current screen stored in 6B in pixel units and integrating the signal.

The smoke amount calculated by the smoke amount integrating unit 28 and the temporal change in the smoke amount calculated by the change amount integrating unit 30 are given to the fire judging unit 32 to judge the presence or absence of a fire.
The fire determination unit 32 is provided with a first determination unit 34 that determines a fire based on the amount of smoke emitted and a second determination unit 36 that determines a fire based on the amount of change in the amount emitted. For the fire determination based on the smoke generation amount in the first determination unit 34, for example, two large and small thresholds are set, and if the threshold is greater than or equal to the larger threshold, it is immediately determined to be a fire. However, when the smoke generation amount is equal to or larger than the smaller threshold value but smaller than the larger threshold value, the second determination unit 36 is requested to make a fire determination based on the change amount of the smoke generation amount. The second determination unit determines the presence or absence of a fire based on the temporal change in the smoke generation amount that exceeds the lower threshold value. For example, if the amount of smoke generated this time is larger than the amount of smoke generated last time and the amount of increase is greater than or equal to a predetermined value, it is determined that there is a fire.

The determination of a fire based on the amount of smoke generated and the amount of change in the amount of smoke generated over time is not limited to a comparative determination based on a fixed threshold value. You may judge it by doing. Further, in the embodiment of FIG. 1, a distance detector 40 is provided, and in this embodiment, the lens movement obtained in the focus control state for the smoke 100 by the autofocus unit 12 of the high sensitivity camera 10. The amount information is captured and the distance to the smoke 100 is detected.

The distance information up to the smoke 100 detected by the distance detection unit 40 is given to the density calculation unit 24, and the size of the detected smoke generation area varies depending on the distance on the screen, and this is corrected. For example, the detected smoking area on the screen when the smoking area is viewed at the reference distance L is S, and the detected smoking area on the screen when the same smoking area is viewed at the doubled distance 2L is S / 2. To do. In this case, the distance is 2L
In order to correct the detected smoke emitting area S / 2 on the screen obtained in step 1 to the size at the reference distance L, the density signal for each pixel is obtained in the corrected detected smoke emitting area with (S / 2) × 2.

In the actual processing, the detected smoking area S is represented by the number of pixels. Therefore, performing the distance correction for each pixel requires thinning processing of pixels and interpolation processing of pixels. Since it is complicated, the smoke density calculation unit 24 obtains a correction coefficient from the smoke reference distance and the detection distance, and sends this correction coefficient together with the density signal to the smoke generation amount integration unit 28 and the change amount integration unit 32 in the subsequent stage, and the integration result is obtained. The correction coefficient may be multiplied according to the distance.

FIG. 3 is a block diagram of an embodiment showing the second embodiment of the present invention. In this embodiment, the area for performing the smoke density calculation processing is fixedly set on the photographing screen. It is characterized by having done. In the embodiment of FIG. 3, a region mask setting unit 42 is provided instead of the smoke generation region detection unit 22 of FIG. The area mask setting unit 42 sets a mask 60 having a predetermined shape and size at a predetermined position on the photographing screen as shown in FIG. As the mask 60,
Since the smoke rises from the bottom to the top due to the hot air flow at the time of a fire, it is desirable to set the slit-shaped mask 60 so that it spreads across the entire width in the center of the screen.

As shown in FIG. 4A, the area mask setting unit 42 sets a mask 60 for determining the area for calculating the smoke amount at a predetermined position on the screen, and at the same time, the high sensitivity camera 10 driven by the camera scanning unit 44. Using the vertical rotation mechanism 46 and the horizontal rotation mechanism 48, feedback control is performed so that the mask 60 is located at the center of the rising smoke 100 as shown in FIG. 4B.

By setting the mask 60 with the feedback control of the camera as described above, the mask 60 can be appropriately set on the portion of the smoke 100 no matter where in the monitoring area smoke from a fire rises. Furthermore, in the embodiment of FIG. 3, since the high-sensitivity camera 10 scans the surveillance space 50 vertically and horizontally, a plurality of light-emissions are provided at positions on the opposite side of the surveillance space 50 facing the high-sensitivity camera 10. The device 14, three light emitting devices 14 in this embodiment, are provided, and the synchronizing device 1 is synchronized with the photographing operation of the high-sensitivity camera 10.
By driving the light emitting device 14 to emit light,
Even if the direction of the is changed, stable backlight illumination can be performed.

Therefore, in the embodiment shown in FIG.
In the setting state of the mask 60 shown in (b), the multi-gradation pixel signal included in the mask 60 is converted into a density signal, and the smoke amount and the change amount of the smoke amount of this time with respect to the previous time are calculated in the same manner as the embodiment of FIG. They will seek and judge the fire. Of course, the distance correction based on the detected distance from the distance detection unit 40 is performed if necessary.

In the embodiment shown in FIG. 3, the setting state for the smoke 100 of the mask 60 shown in FIG. 4B is created by the rotational scanning of the high-sensitivity camera 10.
In the state shown in (a), the high-sensitivity camera 10 may be zoomed up to allow the smoke 100 to enter almost the entire area of the mask 60. Further, when the entire surveillance area can be viewed from the high-sensitivity camera 10, it is not necessary to set the mask 60 by feedback control of the high-sensitivity camera, and image processing for obtaining the smoke generation amount is performed for the mask 60 side set in the camera fixed state. I'll do it.

[0028]

As described above, according to the present invention, since the smoke amount itself of smoke rising in the monitoring space due to the fire is detected, the spot type smoke detection installed on the conventional ceiling surface is detected. It is possible to judge a fire quickly and accurately from the amount of smoke generated without being affected by the size and shape of the monitoring area like a container.

Further, since the smoke amount itself is observed by the image processing, it is possible to clearly distinguish the smoke of the cigarette and the smoke caused by the cooking from the smoke caused by the fire, and the false alarm due to the smoke other than the fire can be surely avoided. High reliability can be obtained. Furthermore, if you can catch it on the screen of the camera, whether smoke is rising, it is flowing along the ceiling surface or along the floor surface, regardless of the state of the smoke, it is possible to reliably generate a fire from the amount of smoke emitted. I can judge.

Further, by displaying the photographed image captured by the camera on the monitor center side together with the result of the fire judgment, it is possible to remotely confirm the abnormal condition of the scene judged to be a fire without going to the scene and operate the device. The reliability above can be greatly improved.

[Brief description of drawings]

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

FIG. 2 is a characteristic diagram showing a correspondence relationship between a multi-tone pixel signal and smoke density in the embodiment of FIG.

FIG. 3 is an explanatory diagram showing a second embodiment of the present invention.

FIG. 4 is an explanatory view showing mask setting control in the embodiment of FIG.

[Explanation of symbols]

10: High sensitivity camera (camera device) 12: Autofocus section 14: Light emitting device 16: Synchronizer 18: AD converter 20A: First frame memory 20B: Second frame memory 22: Smoke area detection unit 24: Smoke concentration calculator 26A: First smoke area frame memory 26B: Second smoke area frame memory 28: Smoke amount integration unit 30: Change amount integration unit 32: Fire judgment section 34: 1st judgment part 36: Second judging unit 38: Subtractor 40: Distance detector 42: Area mask setting section 44: Camera scanning unit 46: Vertical rotation mechanism 48: Horizontal rotation mechanism 50: Surveillance space 60: Mask 100: smoke

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Taku Watanabe             2-1043 Kamiosaki, Shinagawa-ku, Tokyo Ho             Chiki Co., Ltd.

Claims (4)

[Claims] 1. A camera device for imaging a surveillance area and outputting a video signal, a light emitting device installed on a side facing the camera device across a surveillance space, and synchronizing with a photographing operation of the camera device. A synchronizing device for driving the light emitting device to emit light, a frame memory for converting image signals of at least two screens photographed by the camera device into a multi-gradation pixel signal and storing the same, and one screen stored in the frame memory. Smoking area detection unit that detects a smoking area from the multi-gradation pixel signal for each minute, and the smoke density for each pixel is calculated based on the multi-gradation pixel signal in the smoking area detected by the smoke area detection unit. A smoke density calculation unit for storing in the smoke area frame memory, a smoke amount integration section for integrating the smoke density signal for each pixel stored in the smoke area frame memory to calculate the smoke amount, and a smoke amount integration section for calculating Smoking A smoke detection device using image processing, comprising: a fire determination unit that determines a fire based on the amount. 2. A camera device for picking up an image of a surveillance area and outputting a video signal, a light emitting device installed on the side facing the camera means across a surveillance space, and in synchronization with a photographing operation of the camera device. A synchronizing device for driving the light emitting device to emit light, a frame memory for converting image signals of at least two screens captured by the camera device into a multi-gradation pixel signal and storing the same, and an intra-screen stored in the frame memory. A region mask setting unit that sets a mask that determines a predetermined region for performing smoke amount detection processing, and each pixel based on the multi-gradation pixel signal included in the mask region set by the region mask setting unit. Smoke density calculation unit that calculates the smoke density for each and stores it in the smoke area frame memory, and smoke that calculates the smoke amount by integrating the smoke density signal for each pixel stored in the smoke area frame memory An integrating unit, smoke detection apparatus using an image processing, characterized in that it comprises a fire determination section for determining a fire, the based on the amount of smoke that is calculated by emitting smoke quantity integration unit. 3. A smoke detection apparatus using the image processing according to claim 1 or 2, further comprising: a smoke density signal of a current time and a smoke density signal of a previous time stored in the smoke area frame memory. A change amount integration unit for calculating a temporal smoke change amount by calculating the difference for each pixel and integrating is provided, and the fire determination unit is calculated by the smoke amount calculated by the smoke amount integration unit and the change amount integration unit. A smoke detection device using image processing, which is characterized by determining a fire based on the amount of change over time. 4. A smoke detecting apparatus using the image processing according to claim 1 or 2, further comprising a distance detecting section for detecting a distance to a smoke emitting area photographed by the camera device, The smoke detection device using image processing, wherein the calculation unit calculates the smoke density after correcting the size of the smoke image according to the distance.