JP2001023055A - Flame detector and flame detection method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a flame detector which can highly precisely detect flame and whose reliability is improved by detecting the existing position of the flame in a monitoring area and discriminating the presence or absence of the flame in a specified partial picture corresponding to a detection area judged that the flame exists. SOLUTION: The partial picture of a picture obtained by a picture input means 2 can be mode to correspond to the detection areas of respective infrared sensors in a flame position detection means 1. When flame is detected by the infrared sensor in the flame position detection means 1, only the corresponding partial picture is extracted by an area specification means 3 and a flame discrimination means 4 executes a flame discrimination processing based on a picture processing only on the corresponding partial picture. Thus, the highly precise and complicated picture processing required for the discrimination of the presence or absence of the flame is executed only on the partial picture. Calculation quantity required for flame detection can considerably be reduced and the precision of the flame detection can be improved. Thus, a small and inexpensive flame detector can be obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for detecting fire, flame or fire.

[0002]

2. Description of the Related Art As a conventional apparatus for detecting a fire using an image processing apparatus, for example, there is an apparatus described in Japanese Patent Application Laid-Open No. 10-269468. The main principle of such a conventional apparatus is to capture a flame at the time of a fire by transferring all images captured by an image sensor to a RAM and processing the entire image.

[0003]

However, in such a conventional apparatus, since an entire image is always processed, an expensive processor is required for image processing, or an expensive processor is not used. In such a case, it is necessary to perform simple processing with a small amount of calculation, and as a result, there is a problem that a false report may occur.

Further, when judging whether or not a flame is present only from image information, there is another problem that another light source confusing with the flame, for example, reflected light of sunlight, etc., is erroneously recognized as a flame.

The present invention has been made to solve such a problem, and performs image processing only on an area where a flame is likely to exist in an image acquired by an image sensor. It is an object of the present invention to obtain a flame detection device and a method for improving reliability by reducing the amount of processing and performing highly accurate flame detection using an inexpensive processor and combining a plurality of flame detection information. is there.

[0006]

According to the first aspect of the present invention, there is provided a flame detecting apparatus, comprising: image input means for photographing an image of a monitoring area; flame position detecting means for detecting a position of a flame in the monitoring area; An area specifying means for obtaining, from the image input means, a partial image corresponding to a detection area in which a flame is determined to be present by the detecting means, and a flame determining means for determining the presence or absence of a flame in the partial image specified by the area specifying means. It was made.

According to a second aspect of the present invention, there is provided a flame detecting apparatus, comprising: image input means for photographing an image of a monitoring area; flame determining means for determining presence or absence of a flame for each of a plurality of partial images output by the image input means; Flame position detection means having a plurality of flame detection sensors each having a detection area corresponding to the partial image, detection result integration means for receiving the output of the flame position detection means and the output of the flame determination means and outputting an integrated flame detection result Is provided.

According to a third aspect of the present invention, there is provided a flame detecting apparatus, comprising: image input means for photographing an image of a monitoring area; feature extracting means for extracting a feature amount from an image output from the image input means; And a flame discriminating means for discriminating the presence or absence of a flame in the characteristic region specified by the region specifying means.

According to a fourth aspect of the present invention, there is provided a flame detecting apparatus according to the third aspect, wherein the image inputting means includes a feature extracting means.

According to a fifth aspect of the present invention, in the fire detecting apparatus according to the third or fourth aspect, when the flame is determined by the flame determining means, the presence or absence of the flame in the characteristic region is determined. And a flame position detecting means.

According to a sixth aspect of the present invention, there is provided a flame detecting apparatus, comprising: image input means for photographing an image of a monitoring area; flame position detecting means for detecting a position of a flame in the monitoring area; Then, a plurality of flame region specifying units configured by an area specifying unit that obtains a partial image corresponding to the detected region from the image input unit, and the flame region specifying unit uses a network to specify the partial image specified by the region specifying unit. The flame is then transmitted to the flame discriminating means, and the flame discriminating means discriminates the presence or absence of a flame based on the transmitted partial image.

According to a seventh aspect of the present invention, in the flame detecting apparatus according to any one of the first, second, and sixth aspects, the image input means captures a partial image in the monitoring area. It is composed of an image sensor having a function of outputting.

According to a eighth aspect of the present invention, in the flame detecting device according to any one of the first, second, and sixth aspects, the image input means captures an entire image of the monitoring area, It is composed of an image sensor having a function of outputting a partial image.

According to a ninth aspect of the present invention, in the flame detecting apparatus according to any one of the third to fifth aspects, the image input means outputs a projection value calculated from the obtained image. It is composed of an image sensor having a function.

According to a tenth aspect of the present invention, in the flame detecting device according to any one of the third to fifth aspects, the image input means includes a time of the projection value calculated from the obtained image. The image sensor has a function of outputting a change as image information.

According to an eleventh aspect of the present invention, in the flame detecting device according to any one of the first to tenth aspects, the image input means is constituted by an artificial retinal LSI.

According to a twelfth aspect of the present invention, there is provided a flame detecting apparatus according to any one of the first to eleventh aspects, wherein the flame position detecting means comprises a plurality of flame detecting sensors. is there.

According to a thirteenth aspect of the present invention, in the flame detecting device according to the twelfth aspect, the flame detecting sensor comprises an infrared sensor having a function of detecting infrared rays.

According to a fourteenth aspect of the present invention, there is provided a flame detecting apparatus according to the twelfth aspect, wherein the flame detecting sensors are arranged in an array.

According to a fifteenth aspect of the present invention, there is provided a flame detecting method, comprising: image input means for photographing an image of a monitoring area; and flame determining means for judging whether or not there is a flame in the image photographed by the image input means. The flame determining means extracts a flame candidate region from an image, calculates a variation value of the flame candidate region between different image frames, and determines whether the flame candidate region is a true flame based on the calculated variation value. This is to determine whether or not.

According to a sixteenth aspect of the present invention, in the flame detecting method according to the fifteenth aspect, the variation value is a change rate of the area of the flame candidate region.

According to a seventeenth aspect of the present invention, in the flame detecting method according to the fifteenth aspect, the variation value is a change amount of the center of gravity of the flame candidate area.

[0023] In a flame detecting method according to an eighteenth aspect of the present invention, in the flame detecting method according to the fifteenth aspect, the variation value is a value obtained by dividing a change amount of the center of gravity of the flame candidate area by a square root of the area.

[0024]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiment 1 Hereinafter, Embodiment 1 of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a flame detection device according to Embodiment 1 of the present invention, which is composed of four components: a flame position detection unit 1, an image input unit 2, an area identification unit 3, and a flame determination unit 4. Is done.

The image input means 2 captures an area to be monitored (hereinafter referred to as a monitored area) and outputs an image. As shown in FIG. 2, for example, the flame position detecting means 1 includes a plurality of infrared sensors 11 for detecting infrared rays as flame detecting sensors.
It is configured by dimensional arrangement. Each of the infrared sensors 11 in the flame position detecting means 1 is assigned a corresponding detection area, and has a function of detecting a heat source included in each detection area. FIG. 3 shows a flame position detecting means 1 on a wall 32.
FIG. 4 is a diagram showing that a detection area of a certain infrared sensor 11 in the flame position detection means 1 becomes an area 33 when is installed. Here, the combination of the detection areas of the individual infrared sensors 11 included in the flame position detection means 1 is set such that the flame position detection means 1 and the image input means 2 are equal to the monitoring area photographed by the image input means 2. Install. That is, the image input means 2 is installed so that partial images corresponding to the respective detection areas of the infrared sensors 11 in the flame position detecting means 1 can be taken. Here, the image input means 2 may be one that captures and outputs a partial image, or one that captures an entire image and outputs a partial image.

FIG. 4 shows an image 41 photographed by the image input means 2 and each infrared ray in the flame position detecting means 1 when the flame position detecting means 1 comprises a plurality of infrared sensors 11 of 3 rows and 3 columns. FIG. 3 is a diagram showing a correspondence of a sensor 11.
For example, for one infrared sensor 11a in the flame position detecting means 1, one partial image 41 in the image 41
a corresponds. Other infrared sensors and other partial images in the image 41 can be similarly associated.

In the present embodiment, only the partial image of the monitoring area of the image input means 2 corresponding to the area of the infrared sensor 11 in which the flame is detected by the flame position detecting means 1,
A method of performing a flame determination based on image processing by the flame determination unit 4 will be described. Since the partial image of the image obtained from the image input means 2 can be associated with the detection area of each infrared sensor 11 in the flame position detection means 1, the flame is detected by the infrared sensor 11 in the flame position detection means 1. In such a case, only the corresponding partial image is extracted by the area specifying means 3, and the flame discriminating means 4 performs the flame discriminating process based on the image processing only on the corresponding partial image. As the flame discriminating means 4 based on the image processing, a discriminating means conventionally used may be used, or a ninth embodiment described later.
Therefore, a newly proposed discriminating means may be used in the present application described in the eleventh embodiment.

Further, the image input means 2 may be constructed so as to always carry out photographing, or may be constructed so as to carry out photographing only when the flame is detected by the flame position detecting means 1. In that case, power consumption is saved. Furthermore, it goes without saying that the arrangement of the infrared sensors 11 in the flame position detecting means 1 is not limited to 3 × 3, but can be variously arranged. In the present embodiment,
Although an example in which an infrared sensor array is used as the flame detection sensor as the flame position detection means 1 has been described, the wavelength of the electromagnetic wave mainly generated from the flame varies variously depending on the type of burning substance or over time. As a means for detecting the flame position, another detection means such as an ultraviolet sensor array or a visible light sensor array, for example, an artificial retinal LSI described in a later-described seventh embodiment may be used.

Embodiment 2 In the first embodiment, the flame discrimination is performed by the flame discriminating means 4 only for the partial image in which the flame has been detected by the flame position detecting means 1. The flame is determined from the image of the monitoring area input from step 2, the position of the flame is determined, and then the corresponding infrared sensor 11 in the flame position detecting means 1 is operated to determine the presence or absence of the flame from the output. You may comprise.

Embodiment 3 Hereinafter, a third embodiment of the present invention will be described with reference to the drawings. FIG. 5 is a block diagram showing the configuration of a flame detecting device according to Embodiment 3 of the present invention. The flame detecting device includes a flame position detecting unit 1, an image input unit 2, a flame determining unit 4, and a detection result integrating unit 5. Be composed. The flame position detecting means 1 and the image input means 2 are installed in the same manner as in the first embodiment, and each infrared sensor 11 in the flame position detecting means 1 as shown in FIG.
The partial images in the image 41 captured by the image input means 2 respectively correspond to the detection regions (1) and (2).

The flame discriminating means 4 performs image processing on the image obtained by the image input means 2, detects a fire present in the monitoring area, and outputs a detection result and a detection position. As the flame discriminating means 4 based on the image processing, a discriminating means conventionally used may be used, or a newly proposed discriminating means in the present application described in Embodiments 9 to 11 described later may be used. Is also good.

The detection result integrating means 5 integrates the detection result by the flame position detecting means 1 and the detection result by the flame discriminating means 4 to calculate a final detection result. As an integration method, for example, it is determined that a flame exists only when the flame discriminating unit 4 detects a flame at the same position as the position where the flame is detected by the flame position detecting unit 1.

Embodiment 4 FIG. In the third embodiment, as the detection result integrating means 5, it is determined that the flame exists only when the flame determining means 4 detects the flame at the same position as the position where the flame is detected by the flame position detecting means 1. Although the method was shown, the integration means is not limited to this,
Other integration means may be used.

For example, even if the flame is not detected by the flame discriminating means 4 at a position corresponding to the position where the flame is detected by the flame position detecting means 1, if a flame is detected in an adjacent partial image, it is determined that a flame exists. You may make it determine.

Embodiment 5 FIG. Hereinafter, a fifth embodiment of the present invention will be described with reference to the drawings. FIG. 6 is a block diagram showing a configuration of a flame detection apparatus according to Embodiment 5 of the present invention.
It is composed of four components, an area specifying means 3 and a flame discriminating means 4.

The image input means 2 captures an image of the monitoring area. The feature extracting means 62 is the photographed image 41 of the image input means 2
Calculate the feature value from The region specifying unit 3 estimates an image region where a flame exists from the feature amount calculated by the feature extracting unit 62. The flame determining means 4 determines the presence or absence of a flame in the image area estimated by the area specifying means 3 based on image processing.

In the present embodiment, the monitoring area is a room as shown in FIG.
Is fixedly arranged on the wall 32 or the like, and a flame 73 to be detected exists in the room.

FIG. 8 is an example of an image 41 captured by the image sensor 71. An image 81 is a photographed result of the flame 73, and the background does not include other images. Such an image 41
Can be obtained by, for example, attaching a visible light blocking filter to the image sensor 71.

Next, the operation of extracting the characteristic amount from the input image 41 shown in FIG. 8 by the characteristic extracting means 62 will be specifically described with reference to FIG. FIG. 9 illustrates an example showing a feature amount calculated by the feature extracting unit 62 from the image 41 captured by the image sensor 71. Feature extraction means 6
A case will be described in which the feature amount of the imaging target calculated in 2 is calculated as the sum of pixel values in the column direction and the row direction (hereinafter, referred to as vertical projection and horizontal projection, respectively).

In FIG. 9, reference numeral 102 denotes a horizontal projection;
Is a vertical projection, and 103 and 106 are threshold values for the horizontal projection 102 and the vertical projection 105, respectively. As for the values of the horizontal projection 102, the rightward direction in the figure indicates the magnitude of the values. Regarding the vertical projection 105, the downward direction in the figure indicates the magnitude of the value, and the threshold value uses a preset value. Further, as shown in FIG. 9, a section in which the horizontal projection 102 exceeds the threshold value 103 is defined as a feature section 104 and a vertical projection 1
Assuming that a section in which 05 exceeds the threshold 106 is a characteristic section 107, a characteristic partial image 108 is determined corresponding to the characteristic section 104 and the characteristic section 107. That is, the feature extracting unit 62 extracts the feature section 104 and the feature section 107 as feature amounts, and the region specifying unit 3 further extracts the feature partial image 10
8 is extracted and specified.

The flame discriminating means 4 detects the presence or absence of a flame by performing image processing on the extracted characteristic portion image 108. As the flame discriminating means 4 based on the image processing, a discriminating means conventionally used may be used, or a newly proposed discriminating means in the present application described in Embodiments 9 to 11 described later may be used. Is also good.

Embodiment 6 FIG. Hereinafter, a sixth embodiment of the present invention will be described with reference to the drawings. The configuration of the flame detection device according to the present embodiment is the same as that of the fifth embodiment. In the present embodiment, a description will be given of a method of correctly detecting only a flame when an image 113 caused by a luminous body other than a flame exists in an image 41 obtained by the image input unit 2.

FIG. 10 shows that a flame 73 exists in the monitoring area.
Further, an example will be described in which a feature amount calculated by the feature extracting unit 62 is shown in the case where there is a luminous body generated by the incidence of light or sunlight. , An image 41 obtained by the image input means 2 and a horizontal projection 102 calculated by the feature extracting means 62, and a vertical projection 10
5, threshold 103 for horizontal projection, threshold 1 for vertical projection
06 is the same as FIG. 9, but an image 81 due to the flame and an image 113 due to the illuminant other than the flame are reflected in the image 41. In the figure, the feature section 104 in which the horizontal projection 102 exceeds the threshold 103 and the vertical projection 105
The characteristic section 107 exceeding 6 is obtained by the image 81 due to the flame, and includes the characteristic section 119 and the characteristic section 12.
Numeral 1 is obtained from the image 113 of the luminous body.

The area specifying means 3 selects the area where the flame exists by obtaining the time change of the horizontal projection 102 and the vertical projection 105. Because the flame fluctuates,
A temporal change occurs in the obtained projection. On the other hand, since a high-luminance object caused by a light or the like generally has no fluctuation, no temporal change occurs in the projection.

FIG. 11 shows an example of a temporal change in the horizontal projection of an image due to a flame. FIGS. 11A and 11B show examples of horizontal projections obtained at two successive time points. That is, in FIG. 11A, 131 represents a projection at a certain time, and in FIG. 11B, 133 represents a projection after a certain time. Numerals 132 and 134 denote respective threshold values, and the characteristic interval 1 in FIG.
35 and the characteristic section 136 in FIG. 11 (b) are sections where the value of the projection exceeds the respective threshold value. FIG.
137 in (c) shows the absolute value of the difference between the two projections at each position, and 138 shows a threshold value. From FIG. 11C, in the case of an image due to a flame, a characteristic section 139 is obtained corresponding to a temporal change in horizontal projection.
On the other hand, the characteristic section 139 as described above is not obtained in the image formed by the luminous body. In the above example, the horizontal projection has been described, but the vertical projection can be similarly obtained.

As described above, the area specifying means 3 always stores the projection one time before, and obtains the common part of the section 136 and the section 139, so that only the area specified by the image by the flame can be output.

In response to the above output, the flame discriminating means 4 performs a flame detection process based on image processing from the image of the area where the flame exists. As the flame discriminating means 4 based on the image processing, a discriminating means conventionally used may be used, or a newly proposed discriminating means in the present application described in Embodiments 9 to 11 described later may be used. Is also good.

Further, for the section where the horizontal projection and the vertical projection exceed the threshold value, the center of gravity of the projected value can be obtained by calculating the center of gravity of the projection value. Information on the center of gravity of this area is output to the flame discriminating means 4 together with the characteristic portion image 108 of the flame.

Embodiment 7 FIG. Hereinafter, a seventh embodiment of the present invention will be described. The configuration of the flame detection device according to the seventh embodiment is the same as that of the sixth embodiment. However, in the present embodiment, an artificial retinal LSI is used as an example in which the feature extraction unit 62 is included in the image input unit 2. An example will be described.

The artificial retinal LSI is described in “Applied Physics, Vol. 67,
No.4 (1998) Development and Commercialization of Artificial Retina Chip "
As described in, in addition to the function as an image sensor, it has a function of calculating horizontal projection and vertical projection at high speed.
By using this function of the artificial retinal LSI, it is possible to detect the area where the flame exists based on the projection at high speed.

Further, the artificial retinal LSI has a random access function for extracting only an arbitrary rectangular area in the image, so that an image corresponding to the area obtained based on the projection can be extracted at a high speed.

Embodiment 8 FIG. Hereinafter, the eighth embodiment will be described with reference to the drawings. FIG. 12 is a block diagram showing a configuration of a flame detection device according to the eighth embodiment. The flame region identification unit 1 includes three components, ie, a flame position detection unit 1, an image input unit 2, and a region identification unit 3.
50 and 151, and a flame discriminating means 4.

Here, the image input means 1, the flame position detecting means 2, and the area specifying means 3 included in each of the flame area specifying units have the same configuration as in the first embodiment. The plurality of flame area specifying units are connected to a flame discriminating unit 4 via a network.
Connected to The area estimating means 3 in each flame area specifying unit transfers the partial image corresponding to the area determined to be a flame to the flame detecting means 4. The flame detecting means 4 detects the presence or absence of a flame by performing image processing on the partial image received via the network 144. As the flame discriminating means 4 based on the image processing, a discriminating means conventionally used may be used, or a newly proposed discriminating means in the present application described in Embodiments 9 to 11 described later may be used. Is also good.

In this embodiment, as shown in FIG. 13, the flame area specifying units 150 and 151 are installed on the wall 32 and connected to the computer 145 for calculating the procedure of the flame detecting means 4 via the network 144. Characteristic partial image 1
08 is a network 1 from each flame area specifying unit.
The flame is sent to the computer 145 via 44, and the presence or absence of a flame is determined by the flame detecting means 4 operating on the computer 145.

In this embodiment, only the case where two flame region specifying units are installed has been described. However, the number of flame region specifying units is not limited to this, and even if more units are used. The same effect can be obtained.

In the eighth embodiment, the image input means 1, the flame position detecting means 2, and the area specifying means 3 included in the flame area specifying unit have been described as being the same as those in the first embodiment. The structure described in the embodiment, for example, the structure described in the fifth embodiment or the like may be used.

Embodiment 9 FIG. Next, an example of a flame determining means for determining whether or not a flame exists in an image will be specifically described. FIG. 14 is a flowchart showing an algorithm of a conventional general flame discriminating means for judging whether or not a flame exists in an image, which is disclosed in JP-A-10-269468 or JP-A-8-305980. is there.

Hereinafter, the flowchart of FIG. 14 will be briefly described. Since the flame is a region having a high luminance in the image, only the region including the flame is left by binarizing it with an appropriate threshold (step ST101). Subsequently, the remaining regions are labeled, and the regions present together in the image are set as separate flame candidate regions (step ST102).
Subsequently, a correspondence is established between the previously stored flame candidate area of the previous frame (step ST103). It is assumed that the correspondence relationship is, for example, that if the flame candidate regions of the previous frame and the current frame are overlapped or located at a close position, the flames are caused by the same flame. An area that could not correspond to any flame candidate area in the previous frame is newly registered as being caused by a new flame (step ST1).
04). Next, a characteristic amount is calculated and stored for each flame candidate region (step ST105). The calculated feature amount includes, for example, the area of the flame candidate region, the position of the center of gravity, the shape feature, and the like. Next, it is determined whether or not the determination process has been completed for each flame candidate region, and if not, the following process is performed (step ST106). It is determined whether or not the flame candidate area is an area that exists continuously for a predetermined number of frames (step ST107). If they exist consecutively, a discriminating feature is calculated based on the feature stored during that time (step ST108), and it is judged from the discriminating feature whether the flame candidate area is a true flame. (Step ST109). If the region does not exist continuously for the predetermined number of frames, the process proceeds to the next flame candidate region without calculating the feature amount for determination. Then, when the determination processing for all the flame candidate areas is completed, the determination processing for the image is completed.

As the discriminating feature quantity, it is conceivable to look at a change in the area in order to detect a feature in which the flame fluctuates. Used. However, the standard deviation refers to the fluctuation from the average,
For example, as shown in FIG. 15, when the flame rapidly expands and the area rapidly increases within the time for calculating the standard deviation, the average area is larger than the standard deviation due to the original fluctuation. , The standard deviation becomes larger, and thus the value of the standard deviation becomes larger than that due to the original fluctuation. Therefore, in the present embodiment, it is proposed that the average value AT of the absolute value of the change rate of the area given by the following equation be used as the feature amount instead of the variance.

[0060]

(Equation 1)

Here, T is the number of the frame for which the AT is calculated, and Sx is the area of the flame candidate area at the frame number x. M is the number of data to be averaged, and the area information from the frame number T to the frame number T−M before the M frame is used for calculating the AT. When the calculated AT is within a certain range, it is determined that the flame candidate area is a true flame.

Embodiment 10 FIG. The ninth embodiment
Although the average value of the change rate of the area is used as the feature value in the above, the change in the center of gravity of the flame candidate region may be used as the feature value. Regarding the change of the center of gravity, for example, the standard deviation of the position of the center of gravity is observed in the conventional flame discrimination method disclosed in Japanese Patent Application Laid-Open No. 8-305980. However, in this case as well, just like the case of the fluctuation of the area described in the ninth embodiment, when the position of the center of gravity rapidly shifts, the standard deviation becomes larger than the value caused by the fluctuation of the original center of gravity. There is a problem. Therefore, the present invention proposes to use the average JT of the amount of change in the center of gravity given by the following equation as a feature amount.

[0063]

(Equation 2)

Here, T is the number of the frame for which JT is calculated, and Gx is the barycentric coordinate of the flame candidate area in the frame number x. D (a, b) represents the Euclidean distance between the coordinates a and b. M is the number of data to be averaged, and the center of gravity information from the frame number T to the frame number T−M before the M frame is used for the calculation of JT. Then, when the calculated JT is within a certain range, it is determined that the flame candidate area is a true flame.

Embodiment 11 FIG. In the first embodiment,
At 0, the feature value is the average of the absolute value of the change in the center of gravity. However, since the change in the center of gravity is considered to be larger for a larger flame, the change in the center of gravity given by the following equation is calculated as the square root of the area of the flame candidate area. The average JT 'of the values divided by may be used as the feature amount.

[0066]

(Equation 3)

The meaning of each symbol is the same as in the equations (1) and (2). When the calculated JT ′ is within a certain range, it is determined that the flame candidate area is a true flame.

Embodiment 12 FIG. The ninth embodiment
In the eleventh embodiment, the case where one type of discriminating feature amount is used has been described. However, if all discriminating feature amounts are used, it is determined that a flame is present if all are within a predetermined range. Is also good.

Further, Embodiment 9 to Embodiment 1
In No. 1, the temporal average values of the area change rate, the change in the center of gravity, and the rate of movement of the center of gravity were used as the discriminating feature values, but instead of the average value, other values such as a variance value were used. May be used.

It is to be noted that Embodiments 9 to 1 are used.
1, the flame discriminating means used in the flame detecting device according to the present invention described in the first to eighth embodiments has been described. However, the flame discriminating means is not limited to this, but uses an image. If so, another flame detecting device or flame determining means used in a flame detecting method may be used.

The images photographed by the image input means and used in the first to eleventh embodiments are not limited to images using visible light, but may be any images such as infrared images and ultraviolet images that can photograph a flame. May be a simple image.

[0072]

As described above, according to the first aspect of the present invention, there is provided image input means for photographing an image of a monitoring area, flame position detecting means for detecting the presence of a flame in the monitoring area,
Area specifying means for obtaining a partial image corresponding to a detection area where a flame is determined to be present by the flame position detecting means from the image input means, and flame determining means for determining the presence or absence of a flame in the partial image specified by the area specifying means. As a result, advanced and complicated image processing required to determine the presence or absence of a flame need only be performed on partial images, and the amount of calculation required for flame detection can be greatly reduced. There is an effect that the detection device can be configured.

According to the second aspect of the present invention, there is provided image input means for photographing an image of a monitoring area, flame determining means for determining the presence or absence of a flame for each of a plurality of partial images output by the image input means. Flame position detection means having a plurality of flame detection sensors having detection areas respectively corresponding to a plurality of partial images, detection result integration for receiving an output of the flame position detection means and an output of the flame discrimination means and outputting an integrated flame detection result Because the means is provided, the detection result of the flame detection sensor and the determination result using the image are combined for each partial area of the monitoring area, so that it is not determined that the flame is detected unless the flame is detected in the same partial area. Therefore, there is an effect that the accuracy of flame detection can be improved.

According to the third aspect of the present invention, there is provided an image input means for photographing an image of a monitoring area, a feature extracting means for extracting a feature amount from an image output from the image input means,
Since the apparatus is provided with a region specifying unit that specifies a characteristic region based on the characteristic amount output from the characteristic extracting unit and a flame determining unit that determines whether or not there is a flame in the characteristic region specified by the region specifying unit, the detection of the flame position is performed. There is an effect that high-speed operation can be performed with a small amount of calculation.

According to the fourth aspect of the present invention, since the feature extracting means is included in the image input means, a simple configuration can be achieved.

According to the fifth aspect of the present invention, when a flame is determined by the flame determining means, a flame position detecting means for determining whether or not there is a flame in the characteristic region is further provided. There is an effect that the accuracy of flame detection can be improved.

According to the invention, the flame is detected by the image input means for photographing the image of the monitoring area, the flame position detecting means for detecting the position of the flame in the monitoring area, and the flame position detecting means. Area specifying means for obtaining a partial image corresponding to the detection area determined to be present from the image input means,
The flame region specifying unit transmits a partial image specified by the region specifying unit to the flame determining unit using a network, and the flame determining unit transmits the partial image specified by the region specifying unit. Since the presence / absence is determined, the amount of data transmitted via the network can be reduced, so that more monitoring areas can be monitored using the network having the same data transmission rate. Further, by dividing the flame discriminating means from the flame region specifying unit, there is an effect that the flame region specifying unit can be reduced in size, weight and cost.

According to the seventh aspect of the present invention, since the image input means is constituted by an image sensor having a function of photographing and outputting a partial image in the monitoring area, detection of the flame position can be performed with a small amount of calculation. There is an effect that it can be performed at high speed.

According to the invention described in claim 8, the image input means is constituted by an image sensor having a function of taking an entire image of the monitoring area and outputting a partial image, so that the detection of the flame position is reduced. There is an effect that the calculation can be performed at high speed.

According to the ninth aspect of the present invention, the image input means is constituted by an image sensor having a function of outputting a projection value calculated from an obtained image. The effect is that the amount can be increased at a high speed.

According to the tenth aspect of the present invention,
Since the image input means is constituted by an image sensor having a function of outputting the time change of the projection value calculated from the obtained image as image information, the flame position can be detected at high speed with a small amount of calculation. .

According to the eleventh aspect of the present invention,
Since the image input means is constituted by an artificial retina LSI, there is an effect that the flame position can be detected at a high speed with a small amount of calculation.

In the flame detecting apparatus according to the twelfth aspect, the flame position detecting means comprises a plurality of flame detecting sensors. Only need to be done for
There is an effect that the amount of calculation required for flame detection can be greatly reduced, and a small and inexpensive flame detection device can be configured.

According to the thirteenth aspect of the present invention,
Since the flame detection sensor is composed of an infrared sensor that has a function of detecting infrared rays, advanced and complicated image processing required to determine the presence or absence of a flame only needs to be performed on the partial image. There is an effect that the amount of calculation required can be greatly reduced and, consequently, a small and inexpensive flame detector can be configured.

According to the fourteenth aspect of the present invention,
Since the flame detection sensors are arranged in an array,
It is only necessary to perform advanced and complicated image processing required for determining the presence or absence of a flame on only a partial image, so that the amount of calculation required for flame detection can be significantly reduced, and consequently a small and inexpensive flame detection device can be configured. effective.

According to the fifteenth aspect of the present invention,
Image input means for capturing an image of the monitoring area, and flame determination means for determining whether there is a flame in the image captured by the image input means, wherein the flame determination means extracts a flame candidate area from the image. Then, the variation value of the flame candidate region between different image frames is calculated, and it is determined whether or not the flame candidate region is a true flame based on the calculated variation value. Even when the center of gravity moves rapidly, the original fluctuation component of the flame can be extracted as a feature quantity, and there is an effect that the possibility of erroneous determination can be reduced.

According to the sixteenth aspect of the present invention,
The fluctuation value is based on the rate of change of the area of the flame candidate area, so it is determined whether the flame candidate area is a true flame, so if the area suddenly increases or the center of gravity moves suddenly In addition, there is an effect that the original fluctuation component of the flame can be extracted as the feature amount, and the possibility of erroneous determination can be reduced.

According to the seventeenth aspect,
Since the fluctuation value is determined based on the amount of change in the center of gravity of the flame candidate area, whether or not the flame candidate area is a true flame is determined, so if the area suddenly increases or the center of gravity moves suddenly In addition, there is an effect that the original fluctuation component of the flame can be extracted as the feature amount, and the possibility of erroneous determination can be reduced.

According to the eighteenth aspect of the present invention,
Since the variation value is determined based on a value obtained by dividing the change amount of the center of gravity of the flame candidate area by the square root of the area, it is determined whether or not the flame candidate area is a true flame. In the case where moves rapidly, the original fluctuation component of the flame can be extracted as a feature quantity, and there is an effect that the possibility of erroneous determination can be reduced.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a flame detection device according to a first embodiment of the present invention.

FIG. 2 is a diagram showing an example of a flame position detecting means.

FIG. 3 is a diagram showing a state where a flame position detecting means is installed.

FIG. 4 is a perspective view for explaining the correspondence between each infrared sensor and a partial image.

FIG. 5 is a block diagram showing a configuration of a flame detection device according to a third embodiment of the present invention.

FIG. 6 is a block diagram showing a configuration of a flame detection device according to a fifth embodiment of the present invention.

FIG. 7 is a perspective view showing a monitoring area.

FIG. 8 is a plan view showing an image captured by an image sensor.

FIG. 9 is a conceptual diagram for explaining the operation of a feature extraction unit.

FIG. 10 is a conceptual diagram for explaining the operation of a feature extraction unit.

FIG. 11 is a conceptual diagram for explaining the operation of a feature extraction unit.

FIG. 12 is a block diagram showing a configuration of a flame detection device according to an eighth embodiment.

FIG. 13 is a perspective view showing a configuration of a flame detection device according to an eighth embodiment.

FIG. 14 is a flowchart showing an algorithm of a general flame detection unit using a conventional image.

FIG. 15 is a graph illustrating a relationship between a change in the area of a flame candidate region and a feature amount for determination.

[Explanation of symbols]

1 flame position detecting means, 2 image input means, 3 area specifying means, 4 flame detecting means, 5 detection result integrating means, 1
Reference Signs List 1 infrared sensor, 11a infrared sensor, 41 image, 41a partial image, 62 feature extraction means, 71 image sensor, 73 flame, 81 flame image, 102 horizontal projection, 103 horizontal projection threshold, 104 feature section, 105 vertical Projection, 106 vertical projection threshold,
107 feature section, 108 feature partial image, 112 image by flame, 113 image by illuminant, 119 feature section by illuminant, 121 feature section by illuminant, 131
Projection at some point, 132 Projection threshold, 13
3 Projection after a certain time, 134 Projection threshold, 13
5 feature section, 136 feature section, 137 absolute value of projection difference, 138 threshold for absolute value of projection difference, 139 feature section of absolute value of projection difference, 144
Network, 145 computer, 150 flame area identification unit, 151 flame area identification unit.

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Tetsushi Ikeda 2-3-2 Marunouchi, Chiyoda-ku, Tokyo Mitsubishi Electric Corporation (72) Inventor Kazuo Kuma 2-3-2 Marunouchi, Chiyoda-ku, Tokyo F term (reference) in Mitsubishi Electric Corporation 2G065 AA15 AB02 AB04 AB05 BA14 BA15 BA34 BC11 BC14 BC33 CA01 DA06 3K005 QA02 QA04 QA06 QC01 QC04 SA07 5C085 AA13 BA36 CA21 EA41

Claims (18)

[Claims] 1. An image input means for photographing an image of a monitoring area, a flame position detecting means for detecting a position of a flame in the monitoring area, and a detection area corresponding to a detection area where the flame is detected by the flame position detecting means. A flame detecting apparatus comprising: a region specifying unit that obtains a partial image to be obtained from the image input unit; and a flame determining unit that determines the presence or absence of a flame in the partial image specified by the region specifying unit. 2. An image input means for photographing an image of a monitoring area, a flame judgment means for judging the presence or absence of a flame for each of a plurality of partial images outputted by the image input means, and a detection corresponding to each of the plurality of partial images Flame position detection means having a plurality of flame detection sensors having regions, and detection result integration means for receiving an output of the flame position detection means and an output of the flame determination means and outputting an integrated flame detection result. Flame detector. 3. An image input means for photographing an image of a monitoring area, a feature extraction means for extracting a feature quantity from an image output from the image input means, and a feature area specified by a feature quantity output from the feature extraction means. A flame detecting apparatus comprising: an area specifying unit that performs the operation; and a flame determining unit that determines whether there is a flame in the characteristic area specified by the area specifying unit. 4. A flame detecting apparatus according to claim 3, wherein said image input means includes said feature extracting means. 5. The flame according to claim 3, further comprising: flame position detecting means for judging the presence or absence of a flame in the characteristic region when the flame is judged by the flame judging means. Detection device. 6. An image input means for photographing an image of a monitoring area, a flame position detecting means for detecting a position of a flame in the monitoring area, and a detection area corresponding to a detection area where the flame is determined to be present by the flame position detecting means. A plurality of flame region specifying units each comprising: a region specifying unit that obtains a partial image to be obtained from the image input unit. The flame region specifying unit uses a network to determine the flame of each of the partial images specified by the region specifying unit. A flame detecting device, wherein the flame detecting device determines presence or absence of a flame based on the transmitted partial image. 7. The image sensor according to claim 1, wherein said image input means is an image sensor having a function of photographing and outputting a partial image in said monitoring area. The flame detection device as described in the above. 8. The image sensor according to claim 1, wherein said image input means is an image sensor having a function of capturing an entire image of said monitoring area and outputting a partial image. The flame detection device according to any one of the above. 9. The flame according to claim 3, wherein the image input means is an image sensor having a function of outputting a projection value calculated from an obtained image. Detection device. 10. The image sensor according to claim 3, wherein said image input means is an image sensor having a function of outputting, as image information, a temporal change of a projection value calculated from an obtained image. The flame detection device according to any one of the above. 11. The image input means includes an artificial retinal LSI.
The flame detection device according to any one of claims 1 to 10, wherein 12. The flame detecting apparatus according to claim 1, wherein said flame position detecting means comprises a plurality of flame detecting sensors. 13. The flame detection device according to claim 12, wherein the plurality of flame detection sensors are infrared sensors having a function of detecting infrared rays. 14. The flame detection device according to claim 12, wherein the plurality of flame detection sensors are arranged in an array. 15. An image input means for capturing an image of a monitoring area, and flame determining means for determining whether or not there is a flame in the image captured by the image input means, wherein the flame determining means comprises: Extracting a flame candidate region from the image, calculating a variation value of the flame candidate region between different image frames, and determining whether the flame candidate region is a true flame based on the calculated variation value. A flame detection method comprising: 16. The flame detection method according to claim 15, wherein the variation value is a change rate of an area of the flame candidate region. 17. The flame detection method according to claim 15, wherein the variation value is a change amount of a center of gravity of the flame candidate area. 18. The flame detection method according to claim 15, wherein the variation value is a value obtained by dividing a change amount of a center of gravity of the flame candidate area by a square root of an area.