JPH07254096A - Disaster prevention system - Google Patents

Abstract

PURPOSE:To provide a disaster prevention system for improving the accuracy of discriminating fire and non-fire and preventing the violation of privacy. CONSTITUTION:Based on data detected by a sensor 1, a sensor data processor 2 discriminates the fire and the non-fire. At the time of judging fire generation, the sensor data processor 2 transfers the effect of the fire generation to a camera controller 3 along with the position information of the sensor 1 which generated the data. In the camera controller 3, the direction of a television camera 4 is determined so as to pick up the images of a space area including a place with the possibility of the fire generation based on the position information of the sensor 1 and the image pickup of the space area is started. When the image pickup is started by the television camera 4, the images are displayed at a display device 5.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disaster prevention system installed in a building or the like to discriminate between fire and non-fire.

[0002]

2. Description of the Related Art In a general disaster prevention system, a physical phenomenon associated with a fire is detected by a sensor, and the physical quantity of the physical phenomenon (smoke concentration, temperature, carbon monoxide concentration, etc.) is set as an appropriate threshold value set by the sensor. A signal corresponding to fire or non-fire is output from the detector by comparison and determination, and a fire occurrence is notified based on the output signal from the detector. However, in this type of disaster prevention system, if the threshold value is set so that the sensitivity of the sensor is high, false alarms (triggering when there is no fire) are likely to occur, and if the threshold value is set so that the sensitivity of the sensor is low. There is a problem that a false alarm (non-reporting in case of fire) is likely to occur. That is, there is a problem that the accuracy of distinguishing between fire and non-fire is low.

Therefore, in recent years, a sensor whose output value changes according to the physical quantity (smoke concentration, temperature, carbon monoxide concentration) of a physical phenomenon associated with fire is used, and the output value of the sensor is sampled. It has been proposed that the fire / non-fire be discriminated by treating the time series data as time series data and monitoring the variation of the time series data. Furthermore, instead of monitoring only a single physical phenomenon, time series data for multiple types of physical phenomena are used, and the mode of fluctuation of time series data for multiple types of physical phenomena is comprehensively analyzed by an expert system. It is also proposed to judge whether there is a fire or not (Japanese Patent Application No. 3-84958).

[0004]

However, in reality, there are still cases where it is difficult to distinguish between fire and non-fire (cigarette smoke, steam, etc.) only by changing the physical quantity, and only the physical quantity is detected. There is a limit in determining whether a fire or non-fire has occurred. To solve this problem,
It is conceivable to capture an image of a spatial area to be monitored by a television camera and display the image captured by the television camera on a display device such as a CRT. However, since the disaster prevention system is used in a building or the like and generally has a large spatial area to be monitored, it is necessary to constantly monitor images displayed on a large number of display devices. Becomes very large. Further, since the television camera images not only a fire but also a person existing in a space area where the fire should be monitored, there is a possibility that privacy is infringed if the display device constantly displays the image even when there is no fire.

The present invention is intended to solve the above problems, and it is an object of the present invention to provide a disaster prevention system which has a high accuracy of discrimination between fire and non-fire and which does not cause privacy infringement.

[0006]

According to the invention of claim 1, fire detecting means for detecting a fire based on a physical quantity of a physical phenomenon associated with the fire, and the fire detecting means only when the fire detecting means detects the fire. An image input means for picking up a space area including a detection area of the means, and an image display means arranged at a place different from the detection area of the fire detection means and for displaying an image picked up by the image input means on a screen. Is characterized by.

According to a second aspect of the present invention, in the first aspect of the present invention, the fire detecting means is a composite sensor in which a smoke sensor, a heat sensor and a carbon monoxide sensor are combined and provided in one body. Is characterized in that.

[0008]

According to the structure of the first aspect of the present invention, when a fire is detected by the fire detection means, it is determined that the probability of fire occurrence is high, and the spatial area including the detection area of the fire detection means is used as the image input means. Since the image captured by the image input means is displayed on the image display means, the presence or absence of the fire can be visually confirmed by the image when the possibility of fire occurrence is high, and the image is displayed on the image display means. The observer can check the displayed image to make a final determination of fire / non-fire, and it is possible to prevent the occurrence of false alarm during non-fire. Further, since the image input means images the spatial area including the detection area only when the fire detection means detects the fire occurrence, the image display means displays the image near the detection area only when the possibility of fire occurrence is high. Therefore, since the supervisor only has to look at the image display means only when the image is displayed on the image display means, the burden on the supervisor is greatly reduced. In addition, since the detection area is not displayed on the image display means when there is no fire, privacy infringement does not occur.

According to the second aspect of the present invention, since a composite sensor including a smoke detector, a heat detector, and a carbon monoxide detector is used as the fire detecting means, a single type of physical detector is used. Compared with the case of discriminating between fire and non-fire only by the phenomenon, the image input means can start the imaging in the vicinity of the detection area only when there is a high possibility of fire, further reducing the burden on the observer. In addition, the possibility of privacy infringement will be further reduced.

[0010]

In this embodiment, as shown in FIGS. 1 and 2, a composite sensor capable of detecting smoke concentration, temperature and monoxide concentration is used as the sensor 1. That is, the sensor 1 has a configuration in which the scattered light smoke sensor 11, the heat sensor 12, and the carbon monoxide sensor 13 are provided in one body 10. The scattered light type smoke sensor 11 has a light projecting section and a light receiving section which are arranged so that direct light from the light projecting section does not enter the light receiving section, and near the intersection of the optical axes of the light projecting section and the light receiving section. The smoke density is detected by utilizing the fact that the scattered light entering the light receiving part increases when the smoke particles are introduced into, and the analog signal of the output value corresponding to the smoke density is output. The heat sensor uses a thermistor or a thermocouple and outputs an analog signal having an output value corresponding to the ambient temperature. The carbon monoxide sensor 13 uses a semiconductor gas sensor and carbon monoxide. An analog signal with an output value corresponding to the density is output. The analog signals respectively output from the scattered light smoke sensor 11, the heat sensor 12, and the carbon monoxide sensor 13 are
It is sampled by an A / D converter (not shown) provided in the sensor 1 and converted into a digital signal.

A plurality of sensors 1 having such a structure are provided, and each sensor 1 is connected to the sensor data processing device 2 via a sensor line Ls. Here, the sensor 1 is branched and connected to the sensor line Ls, and each sensor 1 has an address so that the sensor data processing device 2 can determine which sensor 1 received the data. It can be identified. That is, the sensor data processing device 2 receives the smoke concentration, temperature, and carbon monoxide concentration data acquired by each sensor 1 by cyclically designating each sensor 1 by an address.

As shown in FIG. 2, the sensor data processing device 2 uses the information from the sensor 1 as a condition section in order to discriminate between fire and non-fire based on the information from the sensor 1. It is provided with a knowledge base 21 that stores knowledge whose conclusion is a non-fire determination result, and an inference engine 22 that searches for knowledge in the knowledge base 21 based on data obtained from the sensor 1. Here, the inference engine 22 is configured to perform fuzzy inference, obtains the certainty factor by integrating the three types of data from each sensor 1, and determines whether a fire or a fire occurs depending on the magnitude relationship between the certainty factor and a prescribed set value. It is designed to identify non-fire. The knowledge stored in the knowledge base 21 corresponds to, for example, the phenomena shown in FIGS. 4 and 5, and the knowledge in which the fire occurrence is the conclusion is shown in FIG.
As shown in, "the smoke concentration (solid line) rises first, then the carbon monoxide concentration (two-dot chain line) rises, and the smoke concentration begins to fall from the maximum value, and at the same time the temperature (one-dot chain line) rises. It has a condition part that says "Start" In addition, the knowledge that non-fire caused by cigarette smoke is the conclusion, as shown in FIG. 5, "carbon monoxide concentration (two-dot chain line) rises first, then smoke concentration (solid line)
Will rise. ” In the inference engine 22, when the data for each sensor 1 is input, the difference between two data that are adjacent in time series is calculated, and the magnitude relationship between this difference and a preset threshold value is calculated to obtain the data. It is determined whether there is an uptrend or a downtrend and there is no change, and by collating with the above-mentioned knowledge in the knowledge base 21, the confidence factor of fire / non-fire is obtained, and the determination result of fire / non-fire is obtained. Output. As described above, the detector 1 and the detector data processing device 2 constitute a fire detecting means.

The sensor data processing device 2 sends the discrimination result obtained as described above and the information on the position of the sensor 1 which has obtained the discrimination result to the camera control device 3. Based on the information obtained from the sensor data processing device 2, the camera control device 3 gives an instruction as to whether or not to take an image to the television camera 4 which can be directed in any direction by a motor, and TV camera 4 if you do
Control the orientation of. The image captured by the television camera 4 is displayed on the display device 5 as an image display unit arranged in a management room or the like.

That is, as shown in FIG. 2, the camera control device 3 has map information relating to the installation positions of the sensor 1 and the television camera 4, and the information relating to the position of the sensor 1 is used as a condition section, and the detection thereof is performed. A knowledge base 31 that stores knowledge required to image the detection area of the detector 1 that has the conclusion and the position and orientation of the TV camera 4, and a detector that detects a fire transmitted through the detector data processing device 2. The expert system comprises an inference engine 32 that finds the required position and orientation of the television camera 4 by matching the position information of No. 1 with the knowledge of the knowledge base 31.
Further, the television camera 4 is capable of adjusting at least a pan angle, which is a rotational angle of rotation of an axis orthogonal to the mounting surface,
If necessary, a tilt angle that is a tilt angle with respect to the mounting surface can be adjusted.

As shown in FIG. 3, a plurality of sensors 1 are arranged on the ceiling of one room, and one or more TV cameras 4 are arranged on the ceiling of the room in which the sensor 1 is arranged. If a fire occurs in this room, the sensor data processing device 2 determines that a fire has occurred based on the variation of the time-series data from each sensor 1, and the sensor that generated the data to obtain the determination result The camera control device 3 with the position information of 1 together with the determination result
Pass to. The camera control device 3 estimates the location of the fire on the basis of the position information of the sensor 1 (when the fire is detected by the information from one sensor 1, the space including the detection area of the sensor 1 is detected. Presumed to be a region, multiple sensors 1
When a fire occurrence is detected in, it is estimated that the detection areas of the detectors 1 overlap each other), and the TV camera 4 arranged at a position where the fire occurrence location can be imaged is selected, and the fire occurrence location is selected. The orientation of the television camera 4 is calculated so that can be imaged. In this way
When the orientation of the TV camera 4 is obtained, the pan angle (and the tilt angle, if necessary) is instructed to the TV camera 4, and the TV camera 4 is activated to start image capturing. At this time, at the same time, the television camera 4 instructs the display device 5 to display the captured image, and the image captured by the television camera 4 is displayed on the display device 5. In this way, the camera control device 3 and the television camera 4 constitute an image input means.

As described above, when it is determined that a fire is likely to occur based on the information detected by the sensor 1, the television camera 4 captures an image of a space area where a fire is highly likely to occur. The image is displayed on the display device 5. Therefore, the television camera 4 does not operate when there is no fire, an image is not unnecessarily displayed on the display device 5, and privacy infringement is prevented. It suffices to look at the display device 5 only when the image is displayed on the screen, and it is possible to reliably determine whether or not an erroneous report occurs, while significantly reducing the burden as compared with the case of constantly monitoring. In addition, the supervisor can judge the fire / non-fire by displaying the scene on the display device 5 without going to the scene, which also reduces the burden on the supervisor and responds to the occurrence of the fire. It can be done quickly. If the alarm sound is also used to notify the observer when there is a high possibility of fire, it is only necessary to look at the display device 5 when the alarm sound is generated. Can be prevented.

In the above configuration, the sensor data processing device 2 sequentially fetches the data input from each sensor 1 and performs the inference by the inference engine. However, a certain number of data for each sensor 1 is used. Alternatively, the data may be stored in the data buffer as time-series data, the contents of the data buffer may be updated each time data is exchanged with the sensor 1, and the data in the data buffer may be inferred.
Further, each sensor 1 is branched and connected via the sensor line Ls, but each sensor 1 is connected to the sensor data processing device 2 via the sensor line Ls so as to be a star connection. You may In this case, the address for each sensor 1 is not always necessary, and the sensor data processing device 2 may switch which sensor line Ls to receive the data by a multiplexer or the like. In addition, the above configuration does not compensate for lost reports,
By setting the criterion for judging the occurrence of a fire based on the information from the sensor 1 to be slightly low, it is possible to surely prevent the false alarm, and the supervisor can deal with the false alarm by visually checking the display device 5. In this way, both false alarms and false alarms can be prevented. Further, when the disaster prevention system of this embodiment is used in a building or the like, the sensor data processing device 2 and the camera control device 4 for the sensor 1 and the TV camera 4 for each floor or each room according to the scale. Alternatively, one sensor data processing device 2 and one camera control device 4 may be provided for each sensor 1 and television camera 4 in the entire building. As many display devices 2 as the TV cameras 4 may be provided,
Since the display is performed only when a fire is detected, it is generally sufficient to switch the display with a smaller number than the TV cameras 4.

[0018]

According to the first aspect of the present invention, when a fire is detected by the fire detecting means, it is determined that the probability of fire occurrence is high, and a spatial area including the detection area of the fire detecting means is imaged by the image inputting means. At the same time, since the image captured by the image input means is displayed on the image display means, the presence / absence of the fire can be visually confirmed by the image when the possibility of fire occurrence is high. There is an advantage that it is possible for a monitor to confirm the image to make a final judgment of fire / non-fire, and to prevent the occurrence of false alarm during non-fire. In other words, the accuracy of distinguishing between fire and non-fire is improved. Further, since the image input means images the spatial area including the detection area only when the fire detection means detects the fire occurrence, the image display means displays the image near the detection area only when the possibility of fire occurrence is high. Therefore, the supervisor only has to look at the image display means only when the image is displayed on the image display means, and the image display means can recognize the presence or absence of a fire without going to the site. There is an advantage that the burden is significantly reduced. Further, since the detection area is not displayed on the image display means when there is no fire, there is an advantage that privacy infringement does not occur.

The invention of claim 2 is a fire detecting means,
Since a combined detector that combines a smoke detector, a heat detector, and a carbon monoxide detector is used, a fire can occur compared to when a single physical phenomenon is used to distinguish between fire and non-fire. The image input means can start imaging in the vicinity of the detection area only when there is a high possibility that the burden on the monitor is further reduced and the possibility of privacy infringement is further reduced.

[Brief description of drawings]

FIG. 1 is a schematic configuration diagram of an embodiment.

FIG. 2 is a block diagram of a main part of the embodiment.

FIG. 3 is an explanatory diagram showing an arrangement example of a sensor and a television camera in the embodiment.

FIG. 4 is an operation explanatory diagram showing an example of knowledge of the sensor data processing device according to the embodiment.

FIG. 5 is an operation explanatory diagram showing an example of knowledge of the sensor data processing device according to the embodiment.

[Explanation of symbols]

 1 Sensor 2 Sensor Data Processing Device 3 Camera Control Device 4 Television Camera 5 Display Device Ls Sensor Line

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical indication H04N 7/18 D

Claims (2)

[Claims] 1. A fire detection means for detecting a fire based on a physical quantity of a physical phenomenon associated with the fire occurrence, and an image of a space area including a detection area of the fire detection means only when the fire detection means detects the fire occurrence. A disaster prevention system comprising: an image input means and an image display means arranged at a place different from a detection area of the fire detection means and displaying an image captured by the image input means on a screen. 2. The fire detecting means is a composite sensor in which a smoke sensor, a heat sensor and a carbon monoxide sensor are combined and provided in one body. Disaster prevention system.