JP6139982B2 - Fire detection device and fire detection method - Google Patents

Description

The present invention relates to a fire detection device and a fire detection method for detecting smoke from a fire from an image of a monitoring area captured by a monitoring camera.

  Conventionally, various devices and systems have been proposed in which a fire is detected by performing image processing on an image of a monitoring area captured by a monitoring camera.

  In such a fire detection device, early detection of a fire is important from the viewpoint of initial extinguishing and evacuation guidance for the occurrence of a fire.

  For this reason, in the conventional device (Patent Document 1), as a phenomenon caused by smoke from the fire from the image, the transmittance or contrast decreases, the luminance value converges to a specific value, and the luminance distribution range narrows. A decrease in dispersion, a change in the average value of luminance due to smoke, a decrease in the total amount of edges, and an increase in intensity in the low frequency band are derived, and these are comprehensively judged to enable smoke detection.

  By the way, in the smoldering combustion which is often in the early stage of fire, a very small amount of smoke rises, the amount of smoke increases with the passage of time, and finally a smoke layer is generated along the ceiling surface. Detects smoke layers generated on the ceiling.

  As described above, it is important to detect a fire at an early stage when a very small amount of smoke rises. However, in a conventional device for detecting smoke by performing image processing on an image, for example, rising smoke The movement (flow) is detected, but it is difficult to detect unless a sufficient amount of smoke rises to detect this flow, and it becomes a very small amount like a thin line. There was a problem that it could not be detected in the early stage of the fire where the smoke of the fire rose.

In order to solve this problem, the applicant of the present application extracts a linear component of a ridge line generated by smoke from an image of a monitoring area captured by a monitoring camera, and obtains a change in time series with respect to the inclination and the occurrence frequency. Has proposed a fire detection device that detects fires by detecting the very small amount of smoke rising due to smoldering combustion, which occurs frequently in the early stages of fire, by detecting characteristic changes caused by smoke. Application 2013-101883).

JP 2008-046916 A JP-A-7-245757 JP 2010-238028 A

  The actual fire situations vary, and it may not be a smoldering fire, but smoke may be generated rapidly, and it is necessary to respond to this.

  As a method of detecting such rapidly generated smoke by image processing, conventionally, an image of a monitoring area captured in a normal state is divided into areas of a predetermined pixel size, and spatial differentiation of each pixel is performed by performing spatial differentiation processing for each area. The sum of the values (edge amount) is calculated, the region with the largest sum of the spatial differential values (the region with the largest luminance difference) is set as the detection region, the sum of the spatial differential values is set as the reference value, and the set detection region The sum of the spatial differential values is detected and monitored, and if smoke is generated, the sum of the spatial differential values decreases. It is used as a standard for determining whether the object that has entered is smoke.

  However, the method of determining smoke from the ratio of the spatial differential sum is to detect the ratio of the spatial differential sum by setting, for example, an area including a window glass in which the luminance difference in the monitoring image is maximum. However, the appearance position of the area where the luminance difference is maximum in the monitoring image varies depending on the monitoring area, and the detection area is set at a specific position that does not correlate with fire smoke in the monitoring image. In some cases, there is a problem that it cannot be determined that smoke has entered from the ratio of the sum of the spatial differential values unless smoke is diffused over the entire monitoring area.

  Also, depending on the monitoring area, there may be a small brightness difference in the entire monitoring image with a uniform brightness, the total sum of the spatial differential values (edge amount) of the area set as the detection area is small from the beginning, even if smoke enters In some cases, the change of the ratio of the sum of the spatial differential values is small, and it may not be possible to determine whether the object that has entered the detection region is smoke or other factors.

An object of the present invention is to provide a fire detection device and a fire detection method that can detect a fire by detecting the generation of smoke from a mask phenomenon in which the stationary background of the monitoring screen is not visible due to the generation of smoke.

(apparatus)
The present invention provides a fire detection device,
An imaging means for capturing an image of the monitoring area;
Ridge line extraction means for performing edge enhancement processing on the image captured by the imaging means and extracting a ridge line;
A linear component extracting unit that divides the image of the ridge line extracted by the ridge line extracting unit into a plurality of image regions, and extracts a linear component of the ridge line for each image region;
For each image area, time-series change detecting means for detecting a change in time series of the linear component extracted by the linear component extracting means;
A background ridge line area extracting means for extracting, as a background ridge line area, an image area in which the time series change of the linear component detected by the time series detection means becomes a characteristic predetermined time series change due to the background;
For the background ridge line area extracted by the background ridge line extraction means, when the time series change of the linear component detected by the time series change detection means disappears, a smoke candidate determination means for determining the background ridge line area as a smoke candidate area;
A fire judgment means for judging a fire based on a distribution of smoke candidate areas detected by the smoke candidate judgment means;
It is provided with.

(Linear component slope and frequency of occurrence)
The linear component extraction means extracts the slope and occurrence frequency of the linear component for each image area,
The time-series change detecting means obtains the time-series change of the slope and occurrence frequency of the extracted linear component,
The background ridge line area extraction means extracts an image area that is a time-series change of a linear component in which the slope of the straight line component is constant and the occurrence frequency is constant, as a background ridge line area,
The smoke candidate determination means determines that the background ridge line area is a smoke candidate area when the occurrence frequency of all the slopes in the time series change of the linear component detected by the time series change detection means is equal to or less than a preset value. To do.

(Fire judgment by estimating smoke candidate range)
If the fire judging means detects a distribution of three smoke candidate areas at least on the most right side of the monitoring screen, the uppermost part from the left, the uppermost part from the left, and the substantially horizontal center of the lowermost part, the three smoke candidate areas A region having the apex at the top is estimated as the smoke distribution range, and a fire is determined when the number of the smoke distribution range is greater than or equal to a predetermined value.
The smoke candidate determination means monitors a time-series change of the linear component for all the extracted background ridge line regions.

(Method)
The present invention provides a fire detection method,
An image of the monitoring area is captured by the imaging means,
Edge enhancement processing is performed on the image captured by the image capturing unit by the ridge line extracting unit to extract the ridge line,
The image of the ridge line extracted by the ridge line extraction means is divided into a plurality of image areas, the linear component of the ridge line is extracted for each image area by the straight line component extraction means,
For each image area, a time-series change detecting unit detects a time-series change of the linear component extracted by the linear component extracting unit,
An image area in which the time-series change of the linear component detected by the time-series detection means becomes a characteristic predetermined time-series change due to the background is extracted as a background ridge line area by the background ridge line area extraction means,
For the background ridge line area extracted by the background ridge line extracting means, when the time series change of the linear component detected by the time series change detecting means disappears, the smoke candidate determining means determines the background ridge line area as a smoke candidate area,
Based on the distribution of the smoke candidate area detected by the smoke candidate judging means, the fire judging means judges the fire,
It is characterized by that.

The other features of the fire detection method of the present invention are basically the same as those of the above-described fire detection device, and thus the description thereof is omitted.

(Basic effect)
According to the fire detection device and the fire detection method of the present invention, an edge enhancement process is performed on an image captured by an imaging unit to extract a ridge line, and the ridge line image is divided into a plurality of image areas. Extracting the linear component of the time series to detect the change in time series, extracting the image area where the time series change of the linear component becomes a characteristic predetermined time series change due to the background as the background ridge line area, When the time-series change of the linear component detected by the time-series change detection means disappears, the background ridge line area is determined as the smoke candidate area, and the fire is determined based on the distribution of the smoke candidate area. All the areas including the extracted background ridgeline become the monitoring area as the background ridgeline area, and there are many background ridgelines in the ridgeline image extracted from the monitoring image. Therefore, many monitoring areas are distributed over the entire monitoring area. If the time series change of the linear component in the background ridge line area disappears due to the generation of smoke, it is assumed that smoke has entered and is designated as a smoke candidate area. Since it spreads, it is possible to judge a fire reliably by looking at the distribution of smoke candidate areas.

(Effects of slope and frequency of ridge line components)
In addition, the slope and occurrence frequency of the straight line component are extracted for each image area of the ridge line image, and the time series change of the slope and the occurrence frequency of the straight line component is obtained. The characteristic straight line of the background in which the slope is constant and the occurrence frequency is constant. When an image area that is a time-series change of the component is extracted as a background ridge line area, and the occurrence frequency of all the slopes in the time-series change of the linear component is less than or equal to a preset value, the background ridge line area is selected as a smoke candidate Since it is determined to be a region, even if the linear component disappears temporarily due to the passage of people or changes in lighting, the frequency of occurrence of all inclinations in the time-series change of the linear component does not fall below a preset value. As long as it is not judged as a smoke candidate area, it is not judged that the background ridge line masking phenomenon (background concealment phenomenon) due to factors other than smoke accidentally entered smoke, and the fire that catches and spreads smoke that rises rapidly due to the fire Detect To allow the door.

Also, a ridge line that is a spatial differential value is extracted from the background image, and then a straight line component (slope and occurrence frequency) of the ridge line is extracted, and further, a time series change ( occurrence frequency for each slope) of the linear component is detected. Since the background ridge line area is specified, the area where the spatial differential sum (edge amount) is maximized due to the large luminance difference is detected, and the presence of smoke is determined based on the ratio of the spatial differential sum of this area. Compared with the conventional method, if there is even one background ridgeline without directly depending on the luminance difference, it is extracted as a background ridgeline region. For this reason, the background ridgeline region is extracted over the entire monitoring screen. Regardless of the position where smoke blows from the screen, the presence of smoke (smoke candidate area) can be determined from the change when smoke enters the corresponding background edge area.

(Effect of fire judgment that estimated smoke candidate range)
The smoke from the fire rises rapidly and then hits the ceiling, so if multiple smoke candidate areas appear in the monitoring area due to the occurrence of smoke, at least the uppermost and the leftmost part of the monitoring screen In the uppermost and lowermost lateral direction, the center of the three candidate smoke areas is estimated as the smoke distribution range, and a fire is judged when the number of areas in the smoke distribution range is greater than or equal to a predetermined value. This makes it possible to detect the fire reliably by capturing the smoke distribution that rises and spreads rapidly.

Further, depending on the monitoring area, there may be few background ridge lines. In this case, the background ridge line areas exist discretely, and the area that becomes a smoke candidate area when smoke enters decreases, but at least three points If the smoke candidate area appears, the smoke distribution range can be estimated, and a fire can be reliably detected even in a monitoring area with a small background ridgeline including a white wall surface.

Explanatory drawing which showed the monitoring area | region where the fire detection apparatus of this invention was installed Block diagram showing outline of functional configuration of image processing apparatus Explanatory diagram showing image segmentation Explanatory diagram showing the time-series change of the linear component of the background area FIG. 4 is an explanatory diagram showing time-series changes of vectors representing the slope and occurrence frequency of the linear component in FIG. Explanatory drawing showing the time-series change of the linear component of other areas that are the background Explanatory drawing which showed the time series change of the vector showing the inclination of linear component of FIG. 6, and the occurrence frequency Explanatory drawing showing the distribution of the background ridge line area extracted from the ridge line image Explanatory drawing showing a ridgeline image of smoke or spout from fire Explanatory drawing which showed distribution of the smoke candidate area | region judged with respect to the background mask by the smoke of FIG. Explanatory drawing which showed extraction of the vertex of the inverted triangle which determines the smoke distribution range estimated from the distribution of the smoke candidate area | region of FIG. Explanatory drawing which showed the smoke distribution range estimated by the inverted triangle of FIG. Explanatory diagram showing an example of the time series change of the smoke distribution range estimated by the inverted triangle that determines the smoke candidate range Flowchart showing the operation of the image processing apparatus of FIG.

[Outline of fire detection device]
FIG. 1 is an explanatory view showing a monitoring area where a fire detection device according to the present invention is installed. FIG. 1 (A) shows a side view, and FIG. 1 (B) shows a front view seen from a monitoring camera.

  As shown in FIG. 1A, a monitoring camera 10 functioning as an imaging unit is installed in the monitoring area 11, and an image is obtained by imaging the state of the monitoring area shown in FIG.

  In the monitoring area 11, equipment such as a trash can 15, a television 16, a sofa 17, and a clock 18 are placed, and straight lines appear in the vertical and horizontal corners of the room and in the vertical and horizontal directions of the wall. ing.

An image captured by the monitoring camera 10 is transmitted to an image processing apparatus 12 installed in a manager's room or the like via a transmission path, and smoke blown from a fire source such as a garbage can 15 is detected by image processing to detect a fire. Judgment is made, and a fire detection signal is output to the fire alarm equipment 14 to output a fire alarm.

[Detection principle]
The principle of detecting smoke by detecting smoke according to the present invention will be described as follows. As shown in FIG. 1, when the surveillance area 11 is imaged by the surveillance camera 10, a variety of background images appear in the captured image. When smoke from a fire blows up and diffuses, the background is masked by the smoke. The fire is judged based on the presence of smoke based on the background mask phenomenon caused by smoke.

  In order to identify a background masked by smoke and disappearing, in the present invention, edge enhancement processing is performed on an image captured by the monitoring camera 10 to extract a ridgeline, and a ridgeline image is generated. The straight line component of the edge line is extracted for each image area, the change of the extracted straight line component in the time series is detected, and the time series change of the detected straight line component is characteristically determined by the background. The image area that changes in time series is extracted as a background ridge line area and monitored.

  When the background is masked by smoke, the background ridge line area disappears from the time-series change of the linear component, so this is detected and the background ridge line area is determined as the smoke candidate area. Determine fire based on distribution.

The straight line component of the ridge line extracted for each image area is its slope and frequency of occurrence, and the straight line component of the background ridge line that exists fixedly has a constant slope and constant frequency of occurrence. An image area having a time-series change ( frequency of occurrence for each inclination) inherent to the image is extracted as a background ridge line area. When the background image is masked by smoke and becomes difficult to see, the frequency of occurrence of all inclinations in the time-series change of the linear component detected for the background ridge line region corresponding to the background image becomes less than a preset value. Therefore, the background ridge line area is determined as the smoke candidate area.

  In addition, since smoke from a fire blows up from the fire source and spreads sideways on the ceiling side, when multiple smoke candidate areas are judged, at least the uppermost part on the right side of the monitoring screen, the uppermost part on the left side, and the uppermost part. Detects the distribution of three smoke candidate areas at the center in the lower horizontal direction, estimates the area having the three smoke candidate areas as apexes as the smoke distribution area, and the number of areas in the smoke distribution area exceeds a predetermined value In case of a fire is judged. As a result, even if the background ridge line region for monitoring smoke is extremely small, the presence of at least three or more points makes it possible to estimate the smoke distribution supplementary range and determine fire.

  Further, it is possible to determine a fire by capturing the state of smoke diffusion in the monitoring area from a time series change estimated as a smoke distribution range having three smoke candidate areas as apexes.

On the other hand, with respect to factors other than smoke, for example, the background mask phenomenon caused by the movement of people wearing plain clothes etc., if the mask phenomenon is temporary, the slope and frequency of the ridge line component will decrease or become zero. However, since the original inclination and frequency of occurrence are restored again and extraction as a smoke candidate area is not continued, a fire is not erroneously determined.

[Fire detection device]
(Functional configuration of fire detection device)
FIG. 3 is a block diagram schematically showing the functional configuration of the fire detection device according to the present invention. As shown in FIG. 3, the fire detection apparatus includes a monitoring camera 10 and an image processing apparatus 12. The image processing apparatus 12 includes a computer circuit having a CPU, a memory, various input / output ports and the like as hardware. Consists of.

  The image processing device 12 functions as a ridge line extraction unit 32 that functions as a ridge line extraction unit, a linear component extraction unit 34 that functions as a linear component extraction unit, and a time-series change detection unit as functions realized by execution of a program by the CPU. A time series change detection unit 36, a background ridge line region extraction unit 38 that functions as a background ridge line region extraction unit, a smoke candidate determination unit 40 that functions as a smoke candidate region determination unit, and a fire determination unit 42 that functions as a fire determination unit are provided. . The transmission unit 30 uses an appropriate transmission interface that receives image data captured by the monitoring camera 10.

  The monitoring camera 10 functioning as an imaging unit transmits image data of a monitoring area, for example, 30 frames per second as moving image data by transmission control of the transmission unit 30 and stores it in a memory (not shown) provided in the image processing device 12. .

  The ridge line extraction unit 32 generates a ridge line image by extracting a ridge line from an image in frame units stored in the memory. For example, in this case, the ridge line extraction unit 32 applies a Sobel filter, which is one of the edge enhancement processes, to the image. For example, as shown in the ridge line image 20 in FIG. The ridge line including 16a, the sofa ridge line 17a, and the clock ridge line 18a is extracted to form a ridge line image. Note that the ridge line extraction processing by the ridge line extraction unit 32 may be performed for each frame in which a predetermined number of frames are thinned out because of the processing speed, without targeting all frame images.

The Sobel filter is present in the image by multiplying the nine pixel values at the upper left and right with a certain target pixel as the center by multiplying a predetermined coefficient by two coefficient matrices in the horizontal and vertical directions to obtain the sum. a differentiation process that allows detecting the boundary of the region (edges) in which, by applying this ridge extraction unit 32 as shown in FIG. 3, extracts a ridge line from the image.

  The straight line component extraction unit 34 divides the ridge line image 20 extracted by the ridge line extraction unit 32 into a plurality of regions, for example, 64 × 64 pixel regions, as shown by dotted lines in FIG. 3, for example, Hough transform (Hough transform) for each region. (Conversion) to extract the linear component of the ridgeline. The Hough transform is known as a method for extracting a straight line segment in an image. For n points in the image, n curves are obtained on the ρ-θ plane. Of these, m curves are obtained. Are intersected at one point, m points on the image corresponding to the m points are on the same straight line, and thereby a linear component can be extracted.

The time-series change detection unit 36 obtains a time-series change ( occurrence frequency for each inclination) of the slope and occurrence frequency of the straight line component extracted by the Hough transform by the straight line component extraction unit 34. In this case, the time-series change detecting unit 36 stores the inclination and occurrence frequency of the straight line component extracted by the straight line component extracting unit 34 in the memory a predetermined number of times in a first-in first-out (first-in first-out) manner. The process of obtaining the slope and its occurrence frequency as a time series change is repeated.

  The background ridge line area extraction unit 38 extracts, as a background ridge line area, an image area having a characteristic predetermined time series change due to the background ridge line from the time series change of the linear component detected by the time series change detection unit 36.

  FIG. 4 shows a time-series change of the straight line component extracted at times t1 to t4 for four cycles of the thinned frame in the area A11 as the background of FIG. In the area A11, four linear components are constantly present in the upper and lower parts in the background. Therefore, at all times t1 to t4, the linear components existing in the area A11 are generated when θ1 = 0 ° upward. The frequency is four.

  FIG. 5 is an explanatory diagram showing a time-series change of the linear component in the area A11 of FIG. 4, and the linear component is shown by accumulating vectors having a slope θ and the length of occurrence frequency.

  As shown in FIG. 5, the vector B1 becomes (θ1, 1) at time 1 and increases from time t2 to t4 in a cumulative manner with a constant occurrence frequency = 4 according to the time series change.

FIG. 6 shows a time-series change of the straight line components extracted at times t1 to t4 for four cycles of the thinned frame in the area A12 as the background of FIG. In the area A12, one straight line component in the vertical direction and six straight line elements in the horizontal direction are constantly present in the background, and in the two-dimensional coordinates with the origin at the center of the lower side of the area, the right side is θ5 = + 90 ° and the left side Is θ6 = −90 ° and the diagonally upper left is θ7, at all times t1 to t4,
The frequency of occurrence of θ1 is one, the frequency of occurrence of θ5 is two, the frequency of occurrence of θ6 is two, the frequency of occurrence of θ7 is two, and the inclination and the frequency of occurrence are constant and constant.

  FIG. 7 is an explanatory diagram showing a time-series change of the linear component in the area A12 of FIG. 6, and shows the linear component accumulated by a vector having a slope θ and a length of occurrence frequency.

As shown in FIG. 7, at time t1, the vector B1 is (θ1, 1).
Vector B5 is (θ5,2)
The vector B6 is (θ6, 2)
The vector B7 is (θ7,2)
From time t2 to time t4, it increases cumulatively at a constant frequency according to the time series change.

  Thus, the time-series change of the linear component of the background ridge line area detected by the time-series change detection unit 36 is a linear component with a constant slope and a constant frequency of occurrence as shown at time t4 in FIGS. The background ridge line area extraction unit 38 extracts an image area that is a steady pattern (time series change) characteristic of the background ridge line as a background ridge line area.

  FIG. 8 is an explanatory diagram showing the distribution of the background ridge line area extracted from the ridge line image of FIG. 3, and as shown by the hatched area, the dust ridge line 15a, the TV ridge line 16a, the sofa ridge line 17a, the clock ridge line 18a, the room A region including a corner stripe and a wall stripe is extracted as a background ridge line area, and the background ridge line area is monitored for changes due to smoke. On the other hand, the area where the background ridge line does not exist is an empty area indicated by a blank, and is excluded from the monitoring area.

The smoke candidate determination unit 40 of FIG. 2 determines that the background ridge line region is a smoke candidate region when the time series change of the linear component detected by the time series change detection unit 36 disappears for the background ridge line region shown in FIG. To do. That is, when the frequency of occurrence of all inclinations in the time-series change of the linear component detected by the time-series change detection unit 36 is equal to or less than a preset value, the smoke candidate determination unit 40 smokes the background ridge line region. Judged as a candidate area.

  FIG. 9 is a ridge line image 20 when the smoke 24 from the fire is spouted with the trash can shown as the trash can ridge line 15a as a fire source. The unit 40 determines the background edge line area masked by the smoke 24 as shown in the smoke candidate area and the black area in FIG. 10, and generates the smoke candidate area 26.

  The fire determination unit 42 in FIG. 2 determines a fire based on the distribution of the smoke region range 26 by the smoke candidate region in FIG. 10 determined by the smoke candidate determination unit 40. For example, the fire determination unit 42, on the screen where the smoke candidate range 26 of FIG. 10 is generated, is the leftmost and uppermost area A1, the rightmost and uppermost area A2, and the lowermost horizontal area A3. As shown in FIG. 12, the distribution of the three smoke candidate regions is detected as shown in FIG. 11, and a plurality of regions corresponding to a range of a substantially inverted triangle having the three smoke candidate regions A1 to A3 as vertices are shown in FIG. The smoke distribution range 28 is estimated, and a fire is determined when the number of areas of the smoke distribution range 28 is equal to or greater than a predetermined value. In this case, since the number of areas corresponding to the minimum inverted triangle range is 3, the fire determination unit 42 determines that there is a fire in the case of a predetermined threshold area that is, for example, 3 areas or more.

  FIG. 13 is an explanatory diagram showing an example of a time-series change of the three smoke candidate ranges forming an inverted triangle and the smoke distribution range estimated thereby. Time t1 in FIG. 13 is a case where smoke starts to blow up, and the inverted triangle that determines the smoke distribution range is small, but as shown at times t2 and t3, the inverted triangle that determines the smoke distribution range with time elapses. It changes so that it spreads upward, and the smoke distribution range will expand. Therefore, the fire determination unit 42 in FIG. 2 can more reliably determine a fire by observing such a time-series change in which the smoke distribution range expands.

[Fire judgment operation]
FIG. 14 is a flowchart showing a fire detection operation by the image processing apparatus of FIG.

In FIG. 14, the image processing apparatus 12 acquires an image of a monitoring area captured at, for example, 30 frames / second as a moving image by the monitoring camera 10 in step S1 (hereinafter “step” is omitted), and stores the acquired image in a memory. Then, a ridge line is extracted from the image by applying a Sobel filter by the ridge line detection unit 32 .

Subsequently, in S3, the image of the ridge line is divided into a plurality of regions by the straight line component extraction unit 34, and the Hough transform is performed for each region in S4 to extract the straight line component of the ridge line. Then, the process proceeds to S5 and the time series change detection unit 36 To obtain a time-series change due to the slope and occurrence frequency of the straight line component extracted in S3, and in S6, the background region extraction unit 38 determines the characteristic of the background from the time-series change of the slope and occurrence frequency of the straight line component for each region. An area having a predetermined time-series change that becomes the inclination and occurrence frequency of the straight line component is extracted as a background ridge line area.

  Subsequently, in S7, the smoke candidate determination unit 40 extracts the area where the disappearance of the background ridge line is detected as the smoke candidate region, and in S8, the fire determination unit 42 determines the fire from the distribution of the smoke candidate region, As a result, when a fire is determined in S9, a fire detection signal is output to the fire alarm facility in S10 to output a fire alarm. On the other hand, if no fire is detected in S9, the process returns to S1 and the same processing is repeated.

[Modification of the present invention]
(Ridge line extraction)
In the above embodiment, the sobel filter is applied to the image to extract the edge of the smoke, but an appropriate filter used for edge enhancement such as a Previtt filter is applied. Also good.

(Linear component extraction)
In the above embodiment, the smoke ridge line is extracted by applying the Hough transform. However, a processing method for extracting a linear component from an image such as Line Segment Detector (LSD) may be applied.

(Image processing device)
In the above embodiment, the surveillance camera and the image processing apparatus are separately arranged and connected by a transmission path, but an apparatus in which both are integrated may be used.

The present invention is not limited to the above-described embodiment, includes appropriate modifications without impairing the object and advantages thereof, and is not limited by the numerical values shown in the above-described embodiment.

10: Surveillance camera 11: Monitoring area 12: Image processing device 14: Fire alarm equipment 15: Garbage bin 16: Television 17: Sofa 18: Clock 20: Edge image 24: Smoke 26: Smoke candidate range 28: Smoke distribution range 30: Transmission unit 32: ridge line extraction unit 34: straight line component extraction unit 36: time-series change detection unit 38: background ridge line region extraction unit 40: smoke candidate determination unit 42: fire determination unit

Claims (8)

An imaging means for capturing an image of the monitoring area;
A ridge line extracting unit that performs edge enhancement processing on the image captured by the imaging unit and extracts a ridge line;
A linear component extraction unit that divides the image of the ridge line extracted by the ridge line extraction unit into a plurality of image regions, and extracts a linear component of the ridge line for each image region;
Time-series change detecting means for detecting a change in time series of the linear component extracted by the linear component extracting means for each image region;
A background ridge line area extracting means for extracting, as a background ridge line area, an image area in which the time series change of the linear component detected by the time series detection means becomes a characteristic predetermined time series change due to the background;
For the background ridge line area extracted by the background ridge line extraction means, when the time series change of the linear component detected by the time series change detection means disappears, a smoke candidate determination means for determining the background ridge line area as a smoke candidate area;
Fire determination means for determining a fire based on the distribution of smoke candidate areas detected by the smoke candidate determination means;
A fire detection device comprising:
In the fire detection device according to claim 1,
The linear component extraction means extracts the slope and occurrence frequency of the linear component for each image area,
The time series change detection means obtains the time series change of the slope and occurrence frequency of the extracted linear component,
The background ridge line area extraction unit extracts an image area that is a time-series change of a linear component having a constant inclination and a constant occurrence frequency as a background ridge line area,
The smoke candidate determination means determines the background ridge line area as the smoke candidate area when the occurrence frequency of all the slopes in the time series change of the linear component detected by the time series change detection means is equal to or less than a preset value. To judge,
A fire detection device characterized by that.
In the fire detection device according to claim 1,
If the fire determination means detects a distribution of at least three smoke candidate areas at the uppermost position on the rightmost side of the monitoring screen, the uppermost position from the leftmost position, and the approximate middle in the lowest horizontal direction, A fire detection device, wherein a region having a candidate region as a vertex is estimated as a smoke distribution range, and a fire is determined when the number of regions in the smoke distribution range is equal to or greater than a predetermined value.
In the fire detection device according to claim 1,
The smoke candidate judging means monitors a time-series change of the linear component for all the extracted background ridgeline regions.
An image of the monitoring area is captured by the imaging means,
The image picked up by the image pickup means is subjected to edge enhancement processing by a ridge line extraction means to extract a ridge line,
The image of the ridge line extracted by the ridge line extraction unit is divided into a plurality of image regions, and the linear component of the ridge line is extracted for each image region by the linear component extraction unit,
For each of the image areas, a time-series change detecting unit detects a change in the time-series of the linear component extracted by the linear component extracting unit,
An image region in which the time-series change of the straight line component detected by the time-series detection unit becomes a characteristic predetermined time-series change due to the background is extracted as a background ridge line region by the background ridge line region extraction unit,
For the background ridge line area extracted by the background ridge line extraction means, when the time series change of the linear component detected by the time series change detection means disappears, the smoke candidate determination means determines the background ridge line area as a smoke candidate area,
Determining a fire by the fire determination means based on the distribution of the smoke candidate area detected by the smoke candidate determination means;
A fire detection method characterized by that.
In the fire detection method according to claim 5 ,
The linear component extraction means extracts the slope and occurrence frequency of the linear component for each image area,
The time series change detecting means detects a time series change of the slope and occurrence frequency of the extracted linear component,
The background ridge line area extraction means extracts a time-series change image area in which the slope of the straight line component is constant and the occurrence frequency is constant, as a background ridge line area,
The smoke candidate determination means determines the background ridge line area as the smoke candidate area when the occurrence frequency of all the slopes in the time series change of the linear component detected by the time series change detection means is equal to or less than a preset value. To judge,
A fire detection method characterized by that.
In the fire detection method according to claim 5 ,
The fire determination means detects a distribution of three smoke candidate areas at least on the most right side of the monitoring screen, on the uppermost part, on the farthest left and on the uppermost part, and at the substantially central part in the lowermost lateral direction. A fire detection method characterized by estimating a substantially inverted triangular area having an area as a vertex as a smoke distribution range, and determining a fire when the number of areas in the smoke distribution range is equal to or greater than a predetermined value.
In the fire detection method according to claim 5,
The smoke candidate judging means monitors a time-series change of the linear component for all the extracted background ridgeline regions.

Images