EP2264677B1

EP2264677B1 - Method for fire prevention and/or detection, and monitoring system and computer product thereof

Info

Publication number: EP2264677B1
Application number: EP10165726A
Authority: EP
Inventors: Giorgio Pelosio
Original assignee: Teletron Euroricherche SRL
Current assignee: Teletron Euroricherche SRL
Priority date: 2009-06-17
Filing date: 2010-06-11
Publication date: 2012-05-30
Anticipated expiration: 2030-06-11
Also published as: ITTO20090459A1; EP2264677A1; IT1394450B1

Description

The present invention relates to a method for fire prevention and/or detection according to the preamble of claim 1. Such method is disclosed in US 2007 000317 .
In addition, the present invention also relates to a monitoring system and a computer product adapted to implement said method.
It is known in the art that the electromagnetic spectrum (or EM spectrum) is the interval of all possible frequencies of radiations, which are electromagnetic waves characterized by a wavelength and a frequency; since wavelength and frequency of a radiation are inversely proportional to each other, the shorter the wavelength the higher the frequency, and thus the energy.
Human beings can perceive with their eyes wavelengths in the range of 380 to 760 nanometres (nm), which are called "visible light".
Shorter wavelengths correspond to ultraviolet rays, X-rays and gamma rays, all of which have frequencies higher than visible light, and therefore more energy than the latter. On the contrary, the wavelengths of radio waves, microwaves and infrared radiations are longer than visible light, and therefore they carry less energy.
Monitoring systems for fire prevention and/or detection are known in the art which comprise automatic panoramic shooting means, thus acquiring images supplied by a plurality of video cameras and assembling said images together in order to immediately provide the operator with a clear global view of the monitored scenario.
More specifically, the fire prevention monitoring systems known in the art employ a plurality of video cameras, which in particular include:

at least a first video camera operating within the visible light range, in particular a Pan Tilt Zoom (PTZ) video camera, to provide the user with a plurality of data and a global view of the surrounding scenario;
at least a second video camera operating within the infrared radiation range.

Said first and second video cameras usually perform a 360° rotation in order to attain a global view of the area to be monitored.
In addition, the fire prevention monitoring systems known in the art comprise:

environmental detection means for detecting some parameters relating to the area to be monitored;
a control unit adapted to receive and analyse the data coming from said plurality of video cameras and from said environmental detection means in order to detect a fire. Consequently, the control unit of known systems performs the following tasks:
it controls the traverse for positioning said plurality of video cameras;
it takes a snapshot by means of the second video camera operating within the infrared radiation range;
it analyses said snapshot, in particular through suitable filters and algorithms, in order to detect any danger and/or starting fire;
it saves the snapshot and pastes it to the previous images in order to create a global view;
it controls the traverse to a next position.

However, it has been observed that the fire prevention monitoring systems known in the art imply a number of drawbacks, in that they are not capable of preserving a high analysis quality as the climatic conditions of the surrounding scenario and of the monitored area change.
In particular, the fire prevention monitoring systems known in the art are not suited to detecting a fire in critical visibility conditions, e.g. when it is dark, rainy, hazy, foggy, etc.
A further drawback suffered by the fire prevention monitoring systems known in the art is that they cannot monitor in a detailed manner the state of the vegetation in the monitored area.
In this frame, it is the main object of the present invention to overcome the above-mentioned drawbacks by providing a fire prevention and/or detection method which ensures a high analysis quality even when the climatic conditions of the surrounding scenario change.
It is another object of the present invention to provide a fire prevention and/or detection method which is suited to detecting a fire even in critical visibility conditions, e.g. when it is dark, rainy, hazy, foggy, etc.
It is a further object of the present invention to provide a fire prevention and/or detection method which can monitor in a detailed manner the state of the vegetation in the surrounding scenario.
Said objects are achieved by the present invention through a fire prevention and/or detection method incorporating the features set out in the appended claims, which are intended as an integral part of the present description.
The present invention also relates to a fire prevention monitoring system as well as to a computer product which can be loaded into a memory of a control unit of the monitoring system, comprising software code portions for implementing said method when the product is executed in the control unit.
Further objects, features and advantages of the present invention will become apparent from the following detailed description and from the annexed drawings, which are supplied by way of an explicative and non-limiting example, wherein:

Fig. 1 is a schematic view of a fire prevention monitoring system according to the present invention;
Fig. 2 is a block diagram of a method for fire prevention and/or detection according to the present invention.

In the annexed drawings, reference numeral 1 designates as a whole a fire prevention monitoring system according to the present invention.
The monitoring system 1 comprises a plurality of video cameras, indicated as a whole by reference numeral 10. In particular, said plurality of video cameras 10 is of the Pan Tilt Zoom (PTZ) type and performs a non-continuous 360° rotation.
According to the present invention, said plurality of video cameras 10 comprises:

at least a first video camera 11 operating within the near infrared range (NIR);
at least a second video camera 12 operating within the thermal infrared range (FIR);
at least a third video camera 13 operating within the ultraviolet range (UV). Preferably, the monitoring system 1 according to the present invention also comprises at least a fourth video camera 14 operating within the visible light range, in order to provide the user with a plurality of data and a global view of the area to be monitored.

Furthermore, the monitoring system 1 according to the present invention comprises:

environmental detection means 20 for detecting some parameters relating to the area to be monitored;
a control unit 30 adapted to receive and analyse the data coming from said plurality of video cameras 10 and from said environmental detection means 20.

In accordance with the present invention, said control unit 30 implements the fire prevention and/or detection method according to the present invention by carrying out the following steps:

a) it controls the traverse for positioning said plurality of video cameras 10,
b) it selects the best performing video camera 11, 12, 13 among said plurality of video cameras 10 on the basis of an entropy value of an image obtained by each video camera 11, 12, 13 and as a function of the data detected by said detection means 20.

As a result, the monitoring system 1 according to the present invention allows to carry out an analysis of the territory within different ranges of the electromagnetic spectrum, so that said analysis can be adapted to the different climatic conditions of the area to be monitored.
As a matter of fact, the infrared range analysis varies much depending on the degree of humidity in the monitored area, i.e. of the degree of transparency of the air in said area.
In turn, the air transparency degree is strongly affected by climatic conditions (in particular, degree of humidity and temperature). The quality of the area analysis performed by using a video camera operating in a certain wavelength may vary considerably; for example, ultraviolet analyses (10 nm - 0.4 µm) are more detailed, but at the same time they are more sensitive to transparency than near IR analyses (0.7 - 1.3 µm), which may be more accurate in the presence of greater atmospheric opacity.
Consequently, the use of a plurality of video cameras 10 operating within different ranges of the electromagnetic spectrum allows to attain optimal fire detection and prevention whatever the environmental condition of the area to be monitored.
In accordance with the present invention, said step b) of selecting a video camera 11, 12, 13 is implemented according to a choice made within a database 31 stored in a memory 32 of said control unit 30, said database 31 concerning the existing relationship between the entropy of the images of said video cameras 11, 12, 13 and the data detected by said detection means 20.
As known, according to the image processing theory, entropy is that parameter which estimates the quantity of information contained in a certain image. The less significant data is present in an image, the closer to zero is the maximum entropy value obtained therefrom; on the contrary, the higher the quantity of significant data contained in an image, the higher the entropy value thereof.
In particular, the method according to the present invention comprises a self-calibration step for automatically adapting to the environmental conditions detected by the detection means 20.
In particular, during said self-calibration step, the control unit 30 implements the following steps at each traverse for positioning said plurality of video cameras 10:

it calculates an entropy value for each image (NIR, FIR, UV) received from said video cameras 11, 12, 13 as a function of the data detected by said detection means 20;
it indicates which image type (NIR, FIR, UV) received from said video cameras 11, 12, 13 has the highest entropy value as a function of said data detected by said detection means 20;
it stores into said database 31 an indication about which video camera 11, 12, 13 among said plurality of video cameras 10 must be selected in the presence of said data detected by the detection means 20.

Thanks to said self-calibration step, the system adapts itself automatically to the environmental conditions without having to calculate an entropy value for each image (NIR, FIR, UV) received from said video cameras 11, 12, 13.
In fact, when the monitoring system 1 is operating normally, the control unit 30 directly uses the image received from said video cameras 11, 12, 13 by associating the data detected by said detection means 20 with the corresponding entropy value; this association is especially quick and advantageous, in that both the data detected by said detection means 20 and the corresponding entropy values have been stored in said database 31 during the self-calibration step.
Preferably, the monitoring system 1 according to the present invention also provides the user with a global view of the monitored area through images supplied by a fourth video camera 14 operating within the visible light range, in particular by associating the images supplied by said fourth video camera 14 with those supplied by a video camera 11, 12, 13 selected among said first 11, second 12 and third 13 video cameras.
Fig. 2 shows a block diagram of the method for fire prevention and/or detection according to the present invention.
At each traverse for positioning said plurality of video cameras, the control unit 30 carries out the following steps:

the datum D_i detected by said detection means 20, in particular corresponding to temperature T_i and humidity H_i at an instant i, is searched for in the database 31 (step 100). If the datum D_i does not exist, it is created (step 101) and added to the database 31 (step 102);
if the datum D_i detected by said detection means 20 has already been associated with a definitive spectrum D_Si, a snapshot is taken based on said definitive spectrum D_Si (step 103), otherwise a calibration is carried out (step 104);
snapshots are taken for each electromagnetic spectrum (UV, NIR and FIR), and, for each of them, a respective entropy value E_UV, E_NIR, E_FIR is calculated (step 105);
the obtained spectrum S_i' having the highest entropy E_MAX is selected (step 106), and it is then verified (step 107) how many times the obtained spectrum S_i' has been selected for that particular datum D_i detected by said detection means 20, in particular corresponding to temperature T_i and humidity H_i. If the number of occurrences R_si is equal to a predetermined maximum number of occurrences R_max, then the obtained spectrum S_i' becomes the definitive spectrum D_Si (step 108), thus ending the calibration for the datum detected by said detection means 20. Otherwise, the relative number of occurrences R_si for the obtained spectrum S_i' is incremented (step 109), and the calibration for the detected datum D_i goes on.

As can be understood from the above description, during the self-calibration step the indication relating to which video camera 11, 12, 13 must be selected in the presence of the data detected by the detection means 20 is only stored into the database when an image type (NIR, FIR, UV) having a maximum entropy value (equivalent to the relative number of occurrences R_si) is received for a determined number of times (equivalent to the maximum number of occurrences R_max set in the system 1).
It is also apparent from the above that said detection means 20 mainly detect data corresponding to a temperature T_i and a degree of humidity H_i at a certain time instant i.
It is clear that the detection means 20 may also detect additional data, such as data pertaining to wind intensity, time of detection, and so on. In such cases, the control unit 30 may increment the maximum number of occurrences R_max in order to adapt the self-calibration step to the increase in the quantity of data to be taken into account; this is essentially equal to saying that the control unit 30 increments the number of times that the picking up of an image type (NIR, FIR, UV) having a maximum entropy value is to be repeated in the presence of said additional data detected by the detection means 20.
It is also plain that the control unit 30 may send the data to a remote centre 40, e.g. via an Internet connection, thus allowing an appropriate fire fighting strategy to be planned.
The advantages of a method for fire prevention and/or detection and a monitoring system thereof according to the present invention are apparent from the above description.
In particular, such advantages consist in that the method for fire prevention and/or detection according to the present invention, as well as the monitoring system thereof, allow a high analysis quality to be preserved as the climatic conditions in the surrounding scenario and in the monitored area change.
In fact, the monitoring system 1 according to the present invention allows to perform an analysis of the territory within different ranges of the electromagnetic spectrum as a function of the data detected by the detection means 20, said data pertaining to temperature, humidity, wind intensity, time of detection, and so on.
Furthermore, the method for fire prevention and/or detection according to the present invention uses that video camera 11, 12, 13 which is most suitable for operating in certain environmental conditions on the basis of the data detected by said detection means 20; this ensures optimization of the fire prevention and detection process whatever the environmental condition in the monitored area.
It follows that the method and system according to the present invention are suited to detecting a fire even in the presence of critical visibility conditions, e.g. when it is dark, rainy, hazy, foggy, etc., as well as to monitoring in a detailed manner the state of the vegetation in the surrounding scenario.
A further advantage of the method and system according to the present invention lies in the fact that, thanks to the execution of a self-calibration step, the method and system according to the present invention adapt themselves automatically to the actual environmental conditions, without having to calculate an entropy value for each image (NIR, FIR, UV) received from said video cameras 11, 12, 13; as a consequence, the monitoring system 1 can directly use the image having the highest entropy, thus reacting very quickly and bringing an unquestionable advantage in terms of fire detection rapidity.
It can therefore be easily understood that the present invention is not limited to the above-described method and system, but may be subject to many modifications, improvements or replacements of equivalent parts and elements without departing from the inventive idea, as clearly specified in the following claims.

Claims

Method for fire prevention and/or detection through a monitoring system (1) comprising:
- a plurality of video cameras (10);

- environmental detection means (20) for detecting some parameters relating to the area to be monitored;

- a control unit (30) adapted to receive and analyse the data coming from said plurality of video cameras (10) and from said environmental detection means (20),
characterized in that
said control unit (30) implements the following steps:
a) it controls the traverse for positioning said plurality of video cameras (10), said plurality of video cameras (10) comprising at least a first video camera (11) operating within the near infrared range (NIR), at least a second video camera (12) operating within the thermal infrared range (FIR), and at least a third video camera (13) operating within the ultraviolet range (UV);

b) it selects the best performing video camera (11, 12, 13) among said plurality of video cameras (10) on the basis of a entropy value of an image obtained by each video camera (11, 12, 13) and as a function of the data detected by said detection means (20);

c) it detects and/or prevents fire by using said selected video camera (11, 12, 13).
Method according to claim 1, characterized in that said step b) of selecting a video camera (11, 12, 13) is implemented according to a choice made within a database (31) stored in a memory (32) of said control unit (30), said database (31) concerning the existing relationship between the entropy of the images of said video cameras (11, 12, 13) and the data detected by said detection means (20).
Method according to claim 1, characterized by comprising a self-calibration step for automatically adapting to the environmental conditions detected by the detection means (20).
Method according to claim 3, characterized in that, during said self-calibration step, the control unit (30) implements the following steps at each traverse for positioning said plurality of video cameras (10):
c) it calculates an entropy value for each image (NIR, FIR, UV) received from said video cameras (11, 12, 13) as a function of the data detected by said detection means (20);

d) it indicates which image type (NIR, FIR, UV) received from said video cameras (11, 12, 13) has the highest entropy value as a function of said data detected by said detection means (20);

e) it stores into said database (31) an indication about which video camera (11, 12, 13) among said plurality of video cameras (10) must be selected in the presence of said data detected by the detection means (20).
Method according to claim 4, characterized in that said step e) of storing into said database (31) the indication of the video camera (11, 12, 13) to be selected only occurs when said step d) of picking an image type (NIR, FIR, UV) having the highest entropy value has been repeated for a predetermined number of times.
Method according to claim 1, characterized in that said detection means (20) detect data corresponding to a temperature (T_i) and a degree of humidity (H_i) at a certain time instant (i).
Method according to claim 1, characterized in that said detection means (20) also detect additional data, in particular relating to wind intensity and time of detection.
Method according to claims 5 and 7, characterized in that said control unit (30) increases the number of times that step d) of picking an image type (NIR, FIR, UV) having the highest entropy value is to be repeated in the presence of said additional data detected by the detection means (20).
Method according to claim 1, characterized by providing a global view of the monitored area through images supplied by a fourth video camera (14) operating within the visible light range, in particular by associating the images supplied by said fourth video camera (14) with those supplied by that video camera (11, 12, 13) which has been selected among said first (11), second (12) and third (13) video cameras.
Method according to claim 1, characterized in that the control unit (30) sends the data to a remote centre (40) in order to allow for planning an adequate fire fighting strategy.
Monitoring system (1) adapted to implement the method according to any of claims 1 to 10.
Computer product which can be loaded into a memory (32) of a control unit (30) of a monitoring system (1), comprising software code portions for implementing the method according to any of claims 1 to 10 when the product is executed in the control unit (30).