Fire detection system, marker position identification method, and program

The fire detection system adjusts camera settings and uses light-emitting markers to overcome placement and imaging condition challenges, ensuring accurate fire area determination by standardizing image sizes, thus improving fire detection precision.

JP2024158488APending Publication Date: 2024-11-08HOCHIKI CORP

Patent Information

Application Number
JP2023073722
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-04-27
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

Existing fire detection systems using image-based markers face challenges in accurately determining the size of a fire area due to marker placement and imaging conditions, such as backlighting, which complicates the identification of the marker's region in the image.

Method used

A fire detection system that adjusts camera settings to reduce overall image brightness and uses light-emitting markers to facilitate accurate determination of marker positions, allowing for correction coefficients to standardize image sizes across different distances and conditions.

Benefits of technology

Enables precise specification of the area where markers are imaged in real space, independent of marker placement or imaging conditions, thereby enhancing the accuracy of fire detection.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024158488000001_ABST
    Figure 2024158488000001_ABST
Patent Text Reader

Abstract

To easily identify a size of a region where a marker disposed in a real space is imaged from an image obtained by imaging the real space regardless of a position where the marker is disposed or an imaging condition.SOLUTION: A fire detection system comprises: a correction setting section for setting a scaling factor for each predetermined region in a picked-up image which is picked up by an imaging section; and a fire detection section for detecting a fire in a target region using the picked-up image which is picked up under a first brightness condition. In a correction mode, the imaging section picks up the target region, in which a light emitter marker is disposed, under a second brightness condition in which brightness is made lower than that of the first brightness condition, and the correction setting section calculates the scaling factor using an image for region correction in which the target region is imaged under the second brightness condition by the imaging section.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to a fire detection system, a marker position specifying method, and a program. [Background technology]

[0002] There is a fire detection system that detects fires using images. For example, Patent Document 1 discloses detecting smoke using images captured in a monitored area. When detecting a fire using images, the range of pixels that can be factors for identifying the fire in the image varies depending on whether the fire breaks out near or far from the image capture position, which may make it difficult to correctly determine the scale of the fire. For this reason, an image is captured in advance in which a marker with a known size is placed in the monitored space to be captured, and the number of pixels corresponding to the marker in the image is measured using the image with the marker. This makes it possible to calculate the size of the real space (real distance) corresponding to a specific pixel range in the image. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2020-057236 A Summary of the Invention [Problem to be solved by the invention]

[0004] However, depending on the position where the marker is placed and the imaging conditions, it may be difficult to identify the area in the image where the marker is captured. For example, when the marker is placed far away from the imaging position or when the color of the entire image changes due to backlighting, it may be difficult to identify the area in the image where the marker is captured. For this reason, there is a problem that it is difficult to identify the area where the marker is captured depending on the position where the marker is placed and the imaging conditions.

[0005] The present invention has been made in consideration of the above circumstances, and its purpose is to provide a fire detection system, a marker position identification method, and a program that can easily identify the size of the area in which a marker placed in real space is imaged from an image of the real space, regardless of the position at which the marker is placed or the imaging conditions. [Means for solving the problem]

[0006] In order to solve the above-mentioned problems, one embodiment of the present invention is a fire detection system that detects a fire using an image in which a target area is captured, the fire detection system comprising: an imaging unit that captures the target area under specific brightness conditions; a correction setting unit that sets a zoom ratio for each predetermined area in the image captured by the imaging unit; and a fire detection unit that detects a fire in the target area using an image that is enlarged or reduced based on the zoom ratio for the image captured by the imaging unit under a first brightness condition. In a correction mode in which the zoom ratio is set by the correction setting unit, the imaging unit captures the target area in which a light-emitting body marker is disposed under a second brightness condition that is brighter than the first brightness condition, and the correction setting unit sets the predetermined area based on the position coordinates of the light-emitting body marker in an area correction image in which the target area is captured by the imaging unit under the second brightness condition, and sets the zoom ratio for the predetermined area calculated based on the size of the light-emitting body marker in the area correction image.

[0007] Moreover, one embodiment of the present invention is a marker position identification method performed by a computer for use in a fire detection system that detects a fire using an image in which a target area is captured, in which an imaging unit images the target area under specific brightness conditions, a correction setting unit sets a zoom ratio for each predetermined area in the image captured by the imaging unit, and a fire detection unit detects a fire in the target area using the image captured by the imaging unit under a first brightness condition. In a correction mode in which the zoom ratio is set by the correction setting unit, the imaging unit images the target area in which a light-emitting marker is disposed under a second brightness condition that is brighter than the first brightness condition, and the correction setting unit sets the predetermined area based on position coordinates of the light-emitting marker in an area correction image in which the target area is captured by the imaging unit under the second brightness condition, and sets the zoom ratio for the predetermined area calculated based on the size of the light-emitting marker in the area correction image.

[0008] Moreover, one embodiment of the present invention is a program for a computer used in a fire detection system that detects a fire using an image in which a target area is captured, which causes the computer to capture the target area under specific brightness conditions, set a zoom ratio for each specified area in the image in which the target area is captured, and detect a fire in the target area using an image that is enlarged or reduced based on the zoom ratio for the image captured under a first brightness condition, and in a correction mode in which the zoom ratio is set but calculated, the target area in which a light-emitting marker is located is captured under a second brightness condition that is brighter than the first brightness condition, and using an area correction image in which the target area is captured under the second brightness condition, the specified area is set based on the position coordinates of the light-emitting marker in the area correction image, and the zoom ratio for the specified area is calculated based on the size of the light-emitting marker in the area correction image. Effect of the Invention

[0009] As described above, according to the present invention, it is possible to easily identify the size of the area in which a marker placed in real space is captured from an image captured of the real space, regardless of the position at which the marker is placed or the imaging conditions. [Brief description of the drawings]

[0010] [Figure 1] 1 is a diagram showing an example of the configuration of a fire detection system 1 to which a fire detection device 10 according to an embodiment is applied. [Diagram 2] 3A to 3C are diagrams showing examples of captured images TG according to the embodiment. [Diagram 3] 3A to 3C are diagrams showing examples of captured images TG according to the embodiment. [Figure 4] 3A to 3C are diagrams showing examples of captured images TG according to the embodiment. [Figure 5A] 11 is a diagram for explaining a method for calculating a correction coefficient according to an embodiment. FIG. [Figure 5B] 11 is a diagram for explaining a method for calculating a correction coefficient according to an embodiment. FIG. [Figure 6A] 11 is a diagram for explaining a method for calculating a correction coefficient according to an embodiment. FIG. [Figure 6B] 11 is a diagram for explaining a method for calculating a correction coefficient according to an embodiment. FIG. [Figure 7] 3 is a flowchart showing a flow of processing performed by the fire detection device 10 of the embodiment. [Figure 8] 3 is a flowchart showing a flow of processing performed by the fire detection device 10 of the embodiment. [Figure 9] FIG. 11 is a diagram for explaining a second modified example of the embodiment. [Figure 10A] FIG. 11 is a diagram for explaining a third modified example of the embodiment. [Figure 10B] FIG. 11 is a diagram for explaining a third modified example of the embodiment. [Figure 11A] FIG. 11 is a diagram for explaining a third modified example of the embodiment. [Figure 11B] FIG. 11 is a diagram for explaining a third modified example of the embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0012] (Fire Detection System 1) The fire detection system 1 is a system that monitors a target area T and detects a fire that has occurred in the target area T. The target area T is an area or object that is the target of monitoring by the fire detection system 1. The target area T may be set arbitrarily. For example, the target area T is a roadway, a tunnel premises, a building, a building premises, etc.

[0013] FIG. 1 is a diagram showing a configuration example of a fire detection system 1 to which a fire detection device 10 according to an embodiment is applied. The fire detection system 1 includes, for example, a camera CA and the fire detection device 10. The camera CA and the fire detection device 10 are connected to be able to communicate with each other. The camera CA and the fire detection device 10 communicate with each other via, for example, a communication network, short-distance wireless communication using a wireless LAN (Local Area Network), or a USB (Universal Serial Bus) cable. The camera CA and the fire detection device 10 may be provided integrally. That is, the camera CA may be built into the fire detection device 10. The camera CA captures an image of a target area T. The camera CA may capture a moving image or a still image. The camera CA outputs the captured image to the fire detection device 10 as a captured image TG. The fire detection device 10 uses the captured image TG to determine whether or not a fire has occurred in the target area T.

[0014] The fire detection system 1 has two modes: a fire detection mode and a correction mode. The fire detection mode is a mode for detecting a fire. In the fire detection mode, a fire is detected by determining whether or not a fire has occurred in the target area T using an image (captured image TG) captured by a camera CA of the target area T.

[0015] In the fire detection mode, it is determined whether or not the captured image TG contains anything that may indicate a fire, such as smoke or flames, and if the captured image TG contains anything that may indicate a fire and the size of that anything that may indicate a fire is equal to or larger than a threshold value, it is determined that a fire has occurred.

[0016] Here, in the captured image TG, due to the law of perspective, things close to the imaging position appear large, and things farther away appear small. Therefore, if something indicating a possible fire occurs near the imaging position, it appears large, and if it occurs far from the imaging position, it appears small. Due to this nature of the image, even if smoke of the same size occurs in the actual target area T, the determination result of whether or not a fire has occurred may differ depending on whether it occurs near or far from the imaging position.

[0017] As a countermeasure, in this embodiment, when detecting a fire in the fire detection mode, the captured image TG is corrected so that objects near and far from the imaging position are of equal size. Specifically, the captured image TG is divided into regions with a certain depth range along the depth direction, and the image of each divided region is enlarged or reduced using a correction coefficient associated with each region. In this way, the captured image TG is corrected so that objects near and far from the imaging position are of equal size.

[0018] In the correction mode, a correction coefficient for each region is calculated, which is used when correcting an image in the fire detection mode. In the correction mode, a marker M is placed at a specific position in the target region T. The marker M is an object that indicates a specified length in real space (e.g., 1 [m], etc.). For example, the marker M is two objects placed at a specified distance apart, or a pole of a specified length with landmark objects provided at both ends.

[0019] The marker M will now be described with reference to Figures 2 to 4. Figures 2 to 4 are diagrams showing examples of captured images TG according to the embodiment.

[0020] 2, the marker M is, for example, spheres B1 and B2 spaced apart by a specified distance. As shown in the example of this figure, in the correction mode, a frame F passing through the spheres B1 and B2 as the marker M is set in the captured image TG, and a correction coefficient is calculated for the area surrounded by the frame F. For example, a coefficient according to the ratio between the number of pixels from the spheres B1 to B2 in the captured image TG and the actual distance (length) from the spheres B1 to B2 in the real space (target area T) is calculated as the correction coefficient.

[0021] On the other hand, FIG. 3 shows an example in which the marker M is placed at a position farther away from the imaging position than the installation position in FIG. 2. In this case, due to perspective, the size of the spheres B1 and B2 indicating the position of the marker M in the captured image TG becomes smaller than that in FIG. 2. In addition, the color of the spheres B1 and B2 changes depending on the imaging conditions such as backlighting or frontlighting. In such a case, it may be difficult to identify the position coordinates where the marker M is captured in the captured image TG.

[0022] To address this issue, this embodiment uses a light-emitting marker M (light-emitting marker M#). By emitting light from the marker M, it becomes possible to easily identify the position coordinates at which the marker M is captured in the captured image TG, regardless of the size of the marker M in the captured image TG and regardless of the imaging conditions, such as backlighting or frontlighting.

[0023] Furthermore, in the present embodiment, in the correction mode, the camera parameters of the camera CA are set so that the entire image in the captured image TG becomes dark, for example, so that the luminance in the captured image TG becomes low, thereby making it possible to easily identify the position coordinates of the light-emitting body marker M# in the captured image TG.

[0024] As a method for darkening the entire image in the captured image TG, for example, it is possible to set the camera parameters so that an aperture operation (iris operation) is performed by narrowing the lens opening and reducing the amount of light entering the camera CA. Alternatively, the camera parameters may be set so as to reduce the gain of an amplifier that amplifies the amount of light received by the camera CA. Also, the camera parameters may be set so as to shorten the accumulation time (exposure time) for accumulating the light received by the camera CA in the imaging element, which corresponds to the shutter beat of the camera CA. Also, these operations, i.e., aperture operation, gain reduction operation, and exposure time reduction operation, may be combined.

[0025] As a method for darkening the entire image in the captured image TG, it is possible to darken the lighting in the target area T, but this is not adopted in this embodiment. This is because it is assumed that the lighting environment of the target area T imaged in the correction mode is the same as the lighting environment of the target area T when detecting a fire. That is, in this embodiment, the settings on the camera CA side are changed, specifically, the camera parameter settings of the camera CA are changed, so that the entire image in the captured image TG becomes dark.

[0026] It is assumed that the imaging position and imaging direction are the same when imaging the target area T in the correction mode and when imaging the target area T when detecting a fire. That is, in the correction mode, the present embodiment does not adopt a method of changing the imaging position and / or imaging direction as a method of darkening the entire image in the captured image TG.

[0027] 4 shows an example of a captured image TG obtained by capturing an image of a target area T in which a light-emitting marker M# is arranged. As shown in the example of this figure, the captured image TG is captured so that the entire image is dark. This makes it possible to easily identify the positions of the spheres B1 and B2, which are the position coordinates of the light-emitting marker M# in the captured image TG.

[0028] (Regarding fire detection device 10) Returning to the explanation of Fig. 1, the fire detection device 10 will now be explained. The fire detection device 10 is a computer, and is realized by, for example, a PC (Personal Computer), a server device, etc. The fire detection device 10 includes, for example, an image acquisition unit 100, a fire detection unit 101, a correction coefficient calculation unit 102, and a correction coefficient storage unit 103.

[0029] The image acquiring unit 100 acquires image information of a captured image TG captured by a camera CA. The image acquiring unit 100 outputs the acquired image information. In a fire detection mode, the image acquiring unit 100 outputs the image information of the captured image TG to a fire detection unit 101. In a correction mode, the image acquiring unit 100 outputs the image information of the captured image TG to a correction coefficient calculation unit 102.

[0030] In the fire detection mode, the fire detection unit 101 acquires a captured image TG from the image acquisition unit 100. The fire detection unit 101 divides the acquired captured image TG into regions in a certain range of depth along the depth direction. The fire detection unit 101 corrects the size of each divided region to be equal in size to the actual distance in the real space (target region T) by enlarging or reducing the image of the region using a correction coefficient associated with each region. For example, the fire detection unit 101 uses a region far from the imaging position in the captured image TG as a reference, and reduces a region close to the imaging position using the correction coefficient of that region. In this way, the size of each region is corrected to be equal in size to the actual distance.

[0031] The fire detection unit 101 judges whether or not something indicating the possibility of a fire, such as smoke or flames, is captured in each corrected region. For example, the fire detection unit 101 judges that something indicating the possibility of a fire is captured when a pixel with a brightness equal to or greater than a threshold exists in each region. When something indicating the possibility of a fire is captured in the captured image TG, the fire detection unit 101 judges that a fire has occurred when the size of the thing indicating the possibility of a fire is equal to or greater than a threshold (for example, 10 [cm]) set for the actual distance. On the other hand, the fire detection unit 101 judges that a fire has not occurred when nothing indicating the possibility of a fire is captured in the captured image TG, and when something indicating the possibility of a fire is captured in the captured image TG but is smaller than the threshold. The fire detection unit 101 outputs the result of judging whether or not a fire has occurred as a detection result of detecting a fire in the target region T.

[0032] In the fire detection mode, the correction coefficient calculation unit 102 acquires the captured image TG from the image acquisition unit 100. Based on the acquired captured image TG, the correction coefficient calculation unit 102 specifies the position coordinates at which the light-emitting body marker M# in the captured image TG is captured.

[0033] The correction coefficient calculation unit 102 sets the camera parameters of the camera CA so that the entire image becomes dark, and causes the camera CA to capture the target area T in which the light-emitting marker M# is placed. The correction coefficient calculation unit 102 acquires image information of the captured image TG thus captured so that the entire image becomes dark, via the image acquisition unit 100. The correction coefficient calculation unit 102 extracts, from the captured image TG, pixels that indicate the light-emitting marker M#, for example, pixels whose brightness is equal to or greater than a threshold value. The correction coefficient calculation unit 102 calculates a correction coefficient based on the position coordinates of the extracted pixels in the captured image TG.

[0034] Here, a method for calculating the correction coefficients by the correction coefficient calculation unit 102 will be described with reference to Fig. 5 (Fig. 5A, Fig. 5B) and Fig. 6 (Fig. 6A, Fig. 6B). Fig. 5 to Fig. 6 are diagrams for explaining the method for calculating the correction coefficients according to the embodiment.

[0035] As shown in FIG. 5A, for example, in the correction mode, a plurality of light-emitting body markers M# (light-emitting body markers M#1 to M#5) are arranged along the depth direction of the target region T. The light-emitting body marker M# is formed by arranging two light-emitting bodies H1 and H2 at a specific distance apart. The correction coefficient calculation unit 102 specifies the position coordinates of each of the plurality of light-emitting body markers M# captured in the captured image TG. For example, the correction coefficient calculation unit 102 specifies the depth direction in the captured image TG, that is, the y-axis direction of the image. The correction coefficient calculation unit 102 extracts two light-emitting bodies H1 and H2 that have the same position coordinates (y coordinate) in the depth direction in the captured image TG and different position coordinates (x coordinate) in the horizontal direction, and defines an illuminant marker M# formed by arranging the two extracted light-emitting bodies H1 and H2.

[0036] As shown in Fig. 5B, the correction coefficient calculation unit 102 uses the specified position coordinates to calculate the number of pixels (number of pixels) corresponding to the distance between the two light emitters H1 and H2 in the light emitter marker M# in the captured image TG for each light emitter marker M#. In the example of this figure, 617 [pixels] are associated with the light emitter marker M1#. In addition, 323 [pixels] are associated with the light emitter marker M#2, 198 [pixels] with the light emitter marker M#3, 130 [pixels] with the light emitter marker M#4, and 86 [pixels] with the light emitter marker M#5.

[0037] As shown in FIG. 6A, the correction coefficient calculation unit 102 extracts, for example, a frame F that passes through each light-emitting marker M# and has a width that is a constant multiple of the length (number of pixels) of each light-emitting marker M# in the image from the captured image TG. For example, the correction coefficient calculation unit 102 extracts a frame F1 from the captured image TG, having a width that is a constant multiple (for example, 1.5 times) of the number of pixels based on the length (617 [pixels]) of the light-emitting marker M#1. Similarly, the correction coefficient calculation unit 102 extracts, from the captured image TG, each of frames F2 to F5, having a width that is a constant multiple (for example, 1.5 times) of the number of pixels based on the length of each of the light-emitting markers M#2 to M#5. The correction coefficient calculation unit 102 calculates a correction coefficient for each area surrounded by the extracted frame F. Furthermore, when extracting each of frames F1 to F5, the correction coefficient calculation unit 102 may extract frame F having a height that is a constant multiple (e.g., three times) the length of the light-emitting marker M#. In this case, if the width and height of frame F are set to a constant multiple of the length of the light-emitting marker M# and the width and height exceed the width of the image, the correction coefficient calculation unit 102 sets the width and height of frame F with the image width as the upper limit.

[0038] As shown in FIG. 6B, the correction coefficient calculation unit 102 calculates a reduction ratio (one example of a magnification / reduction ratio) as a correction coefficient based on the number of pixels (number of pixels) corresponding to each distance of the light-emitting markers M#1 to M#5 in the captured image TG. For example, the correction coefficient calculation unit 102 calculates the reduction ratio based on the length (86 [pixels]) on the image of the light-emitting marker M#5 that is disposed farthest from the imaging position. In this case, the correction coefficient calculation unit 102 calculates the reduction ratio corresponding to the light-emitting marker M#k using the following formula (1). In formula (1), Sn(k) indicates the reduction ratio of the light-emitting marker number k (k=1 to 4). P(k) indicates the length (number of pixels) on the image of the light-emitting marker number k (k=1 to 4).

[0039] Su(k)=P(5) / P(k)×100[%] …Formula (1)

[0040] As shown in Fig. 6B, the correction coefficient calculation unit 102 calculates the reduction ratio corresponding to the light-emitting body marker M#1 to be 13.9[%] using formula (1). Similarly, the correction coefficient calculation unit 102 calculates the reduction ratios corresponding to the light-emitting body markers M#2 to M#5 to be 26.6[%], 34.4[%], 66.2[%], and 100.0[%], respectively. The information shown in Fig. 6B is an example of information stored in the correction coefficient storage unit 103.

[0041] The memory area (including the correction coefficient memory unit 103) in the fire detection device 10 is composed of a storage medium, such as a hard disk drive (HDD), a flash memory, an electrically erasable programmable read only memory (EEPROM), a random access read / write memory (RAM), a read only memory (ROM), or any combination of these storage media.

[0042] In addition, the functional units (image acquisition unit 100, fire detection unit 101, and correction coefficient calculation unit 102) of the fire detection device 10 are realized by causing a CPU (Central Processing Unit) provided as hardware in the fire detection device 10 to execute a program.

[0043] (Flow of processing performed by fire detection device 10) Here, the flow of processing performed by the fire detection device 10 will be described with reference to Fig. 7 and Fig. 8. Fig. 7 and Fig. 8 are flowcharts showing the flow of processing performed by the fire detection device 10 according to the embodiment.

[0044] 7 shows a flow of processing performed by the fire detection device 10 in the correction mode. The fire detection device 10 acquires image information of the captured image TG in the target area T in which the light-emitting marker M# is arranged (step S10). The fire detection device 10 identifies the position coordinates where the light-emitting marker M# is captured in the acquired captured image TG (step S11). The fire detection device 10 calculates the number of pixels corresponding to the length indicated by the light-emitting marker M# in the captured image TG (step S12). The fire detection device 10 calculates a correction coefficient (reduction ratio) using the calculated number of pixels (step S13). The fire detection device 10 stores the calculated correction coefficient (reduction ratio) in the correction coefficient storage unit 103 (step S14).

[0045] FIG. 8 shows a flow of processing performed by the fire detection device 10 in the fire detection mode. The fire detection device 10 acquires image information of a captured image TG in which a target region T is captured (step S20). The fire detection device 10 extracts smoke detection regions from the acquired captured image TG (step S21). The fire detection device 10 divides the captured image TG along the depth direction into regions whose y coordinates are in a specific range, and extracts the divided regions as smoke detection regions. The smoke detection regions are regions in which smoke candidates, that is, things that indicate the possibility of a fire, are detected. The smoke detection regions are, for example, regions surrounded by each of the frames F1 to F5 shown in FIG. 6A. The fire detection device 10 executes processing to detect a fire for each smoke detection region.

[0046] The fire detection device 10 acquires the coordinates of an unprocessed smoke detection area (the processing here is a processing for detecting a fire, which will be described later) (step S22). The coordinates of the smoke detection area here are reference coordinates in the smoke detection area, for example, coordinates located at the lower left of the area. The fire detection device 10 extracts the unprocessed smoke detection area from the captured image TG using the acquired coordinates (step S23). The fire detection device 10 resizes (enlarges or reduces) the extracted smoke detection area using a correction coefficient corresponding to the area (step S24). The fire detection device 10 refers to the correction coefficient storage unit 103 based on the coordinates of the extracted smoke detection area, and acquires a correction coefficient corresponding to the smoke detection area. The fire detection device 10 resizes the smoke detection area using the acquired correction coefficient. The fire detection device 10 stores an image corresponding to the resized smoke detection area in a storage unit (not shown) (step S25). The fire detection device 10 detects a fire by determining whether or not an image indicating the possibility of a fire has been captured in the resized smoke detection area (step S26). The fire detection device 10 determines whether or not there is an unprocessed smoke detection area (step S27), and if there is an unprocessed smoke detection area, returns to step S22 and repeats the process of detecting a fire. If there is no unprocessed smoke detection area, the fire detection device 10 ends the process.

[0047] As described above, the fire detection system 1 of the embodiment includes a camera CA (imaging unit), a fire detection unit 101, and a correction coefficient calculation unit 102. The camera CA captures the target area under a specific brightness condition. The fire detection unit 101 detects a fire in the target area T by using an image TG captured by the imaging unit under a first brightness condition. The correction coefficient calculation unit 102 calculates a correction coefficient for calculating a range of real space (actual distance) in the target area T from the image TG. In a correction mode in which the correction coefficient calculation unit 102 calculates the correction coefficient, a light-emitting body marker M# is placed at a specific position in the target area T. The camera CA captures the target area T under a second brightness condition. The second brightness condition is a condition in which the brightness is lower than that of the first brightness condition, and is a condition in which the camera parameters are set so that the entire image becomes dark. The correction coefficient calculation unit 102 uses a captured image TG (image for area correction) in which the target area T is captured by the camera CA under a second brightness condition to identify the position coordinates of the light-emitting body marker M# in the image for area correction. The correction coefficient calculation unit 102 calculates, as a correction coefficient, a coefficient (e.g., a reduction ratio) for calculating the range (actual distance) of the real space (target area T) corresponding to a specific pixel in the captured image TG, based on the identified position coordinates of the light-emitting body marker M#.

[0048] As a result, in the fire detection system 1 of the embodiment, the camera parameters are set so that the entire image becomes dark, and the position where the light-emitting marker M# is captured in the captured image TG, which captures the target area T in which the light-emitting marker M# is located, can be identified. Therefore, by capturing the captured image TG so that the entire image becomes dark, it becomes easy to identify the position where the light-emitting marker M# is located. Therefore, regardless of the position where the marker is located or the imaging conditions, it becomes easy to identify the size of the area where the marker located in the real space is captured from the image captured of the real space.

[0049] (First Modification of the Embodiment) A first modified example of the embodiment will be described. This modified example differs from the above-described embodiment in that in the correction mode, a composite image is generated using a captured image TG (referred to as a first image) captured with normal camera parameters that do not darken the entire image, the same as in the fire detection mode, and a captured image TG (referred to as a first image) captured so that the entire image is darkened.

[0050] The correction coefficient calculation unit 102 captures a first image and a second image. For example, the correction coefficient calculation unit 102 alternately captures the first image and the second image. The correction coefficient calculation unit 102 identifies the position coordinates of the light-emitting body marker M# using the second image, and generates a composite image by combining a mark indicating the position coordinates of the light-emitting body marker M# with the first image using the identified position coordinates. The composite image generated by the correction coefficient calculation unit 102 is used, for example, by a user who has placed the light-emitting body marker M# in the target area T to determine whether or not there is an error in the position coordinates identified by the correction coefficient calculation unit 102.

[0051] Here, in the first image, depending on the imaging conditions such as backlighting, it may be difficult to identify the position coordinates of the light-emitting marker M# (see FIG. 3). In addition, in the second image, since the entire image is dark, it may be difficult to visually recognize structures other than the light-emitting marker M# (see FIG. 4). To address this issue, in this embodiment, a composite image is generated in which a mark indicating the position coordinates of the light-emitting marker M# is combined with the first image using the position coordinates of the light-emitting marker M# identified using the second image. This makes it easier for the user who has placed the light-emitting marker M# in the target area T to determine whether or not there is an error in the position coordinates identified by the correction coefficient calculation unit 102.

[0052] As described above, in the fire detection system 1 according to the first modification of this embodiment, in the correction mode, the camera CA further captures the target area T under the first brightness condition. The correction coefficient calculation unit 102 identifies the position coordinates of the light-emitting marker M# using the second image (image for area correction). The correction coefficient calculation unit 102 generates a composite image in which the position coordinates of the light-emitting marker M# are indicated in the first image (image captured under the first brightness condition) captured by the camera CA. This makes it easier for the user who has placed the light-emitting marker M# in the target area T to determine whether or not there is an error in the position coordinates identified by the correction coefficient calculation unit 102 in the fire detection system 1 according to the first modification of this embodiment.

[0053] (Modification 2 of the embodiment) A second modification of the embodiment will be described below. This modification differs from the above-described embodiment in that the brightness of the light emitted by the light-emitting marker M# is changed in a specific pattern (frequency).

[0054] Fig. 9 is a diagram for explaining Modification 2 of the embodiment. Fig. 9 shows a schematic diagram of a state in which a plurality of captured images TG are captured by repeatedly capturing images of a target area T in chronological order.

[0055] As shown in Fig. 9, the correction coefficient calculation unit 102 periodically captures an image of the target area T according to the pattern (frequency) of light emitted by the light-emitting marker M#. The correction coefficient calculation unit 102 extracts candidates for the position coordinates of the light-emitting marker M# in one captured image TG (first captured image) of the captured captured images TG. Using the remaining captured images TG, the correction coefficient calculation unit 102 executes a match determination to determine whether or not a pattern (frequency) that changes in time series, calculated from pixel values ​​of pixels corresponding to the candidates for the position coordinates extracted from the first captured image, matches the light emission pattern (frequency) of the light-emitting marker M#. For example, the correction coefficient calculation unit 102 executes the match determination by the method described below.

[0056] First, the correction coefficient calculation unit 102 periodically acquires the brightness (e.g., brightness V in the HSV color space) value of the pixel corresponding to the candidate position coordinate extracted from the first captured image over a predetermined period (e.g., 1 second). The correction coefficient calculation unit 102 calculates the deviation of each of the acquired brightness V values. For example, when 30 frames of images are acquired per second, the brightness V of a specific pixel extracted from each of the 30 images corresponding to the frame images for 1 second is acquired. The correction coefficient calculation unit 102 calculates each deviation as a difference between the average value of the brightness V of each of the acquired 30 points and each value of the 30 points. The correction coefficient calculation unit 102 calculates a correlation value C with the light emission pattern (frequency) of the light-emitting body marker M# for the time series change of each calculated deviation using the following formula (1). In formula (1), C indicates the correlation value, t indicates time, V(t) indicates the deviation of the brightness V at time t, and ω indicates the frequency corresponding to the light emission pattern of the light-emitting body marker M#.

[0057] C=Sqrt(A^2+B^2) …(1) where A = Σ(V(t) × sin(ωt)). B = Σ(V(t) × cos(ωt))

[0058] In formula (1), when the pattern (first frequency) in which the pixel brightness V changes over time matches the light emission pattern (second frequency) of the light-emitting marker M#, the correlation value C becomes large. On the other hand, when the first frequency and the second frequency do not match, the correlation value C becomes small. Using this property, the correction coefficient calculation unit 102 determines that the first frequency and the second frequency match when the correlation value C exceeds a threshold value. On the other hand, when the correlation value C is smaller than the threshold value, the correction coefficient calculation unit 102 determines that the first frequency and the second frequency do not match. When the first frequency and the second frequency match, the correction coefficient calculation unit 102 determines that the pixel extracted from the first captured image is the pixel where the light-emitting marker M# is captured.

[0059] 9, cos(ωt)+isin(ωt) is a schematic representation of the complex number multiplied by V(t) in equation (1). By taking the absolute value of the value obtained by multiplying V(t) by the complex number, the degree of agreement with the frequency ω can be calculated as a correlation value C.

[0060] In addition, in formula (1), V(t) is the deviation of brightness V at time t. Therefore, the correlation value C is different when V(t) is the light-emitting marker M# and when it is another object that blinks at the same frequency as the light-emitting marker M#. Even if the other object blinks at the same frequency as the light-emitting marker M#, if the brightness fluctuation of the other object is smaller than that of the light-emitting marker M#, the correlation value C will not be a large value. Therefore, it is possible to exclude such another object by judgment using the above-mentioned threshold value. Therefore, it is possible to reduce the possibility that the other object will be mistakenly recognized as the light-emitting marker M#.

[0061] As a result of the frequency analysis, the correction coefficient calculation unit 102 identifies pixels where the frequencies of both the pattern in which pixel values ​​change and the pattern in which the brightness of the light-emitting marker M# varies match as pixels corresponding to the position where the light-emitting marker M# is located, and identifies the position coordinates of the identified pixel as the position coordinates where the light-emitting marker M# was imaged.

[0062] In this case, the correction coefficient calculation unit 102 may set the camera parameters so that the entire image becomes dark in the correction mode, as in the above-mentioned embodiment, and perform imaging. Alternatively, imaging may be performed with normal camera parameters. In addition, the correction coefficient calculation unit 102 may be configured to execute a process of specifying the position coordinates at which the light-emitting body marker M# is imaged, using the captured image TG captured with normal camera parameters, and to set the camera parameters so that the entire image becomes dark when the position coordinates cannot be specified.

[0063] As described above, in the fire detection system 1 according to the second modification of the embodiment, in the correction mode, the camera CA captures images of the changes in the light emitted by the light-emitting marker M# in a specific pattern in a time series. The correction coefficient calculation unit 102 uses a plurality of captured images TG (area correction images) captured in a time series by the camera CA to identify the position coordinates of the light-emitting marker M# in those captured images TG. The correction coefficient calculation unit 102 calculates, as a correction coefficient, a coefficient for calculating the range (actual distance) of the real space corresponding to a specific pixel in the captured image TG based on the identified position coordinates of the light-emitting marker M#. As a result, in the fire detection system 1 according to the second modification of the embodiment, the brightness of the light emitted by the light-emitting marker M# is changed, and the position coordinates of a pixel whose pixel value (for example, luminance) changes in response to the change can be identified as the position coordinates of the light-emitting marker M#. This makes it possible to more accurately identify the position coordinates of the light-emitting marker M#.

[0064] In addition, in the fire detection system 1 according to the second modification of the embodiment, the camera CA may image the target area T under a specific brightness condition, image the target area T under a first brightness condition in a fire detection mode (a fire detection mode in which a fire is detected by the fire detection unit 101), and image the target area T under a second brightness condition in a correction mode. This makes it possible to more easily and accurately identify the position coordinates of the light-emitting body marker M#.

[0065] Furthermore, in the fire detection system 1 according to the second modification of the embodiment, the correction coefficient calculation unit 102 extracts candidates for the position coordinates of the light-emitting marker M# using a first captured image (first area correction image) of multiple captured images TG (area correction images) captured in time series by the camera CA. The correction coefficient calculation unit 102 performs frequency analysis based on the time-series changes in pixel values ​​of the light-emitting pixels corresponding to the extracted candidates for the position coordinates, and identifies the position coordinates of the light-emitting marker M#. This makes it possible for the fire detection system 1 according to the second modification of the embodiment to more easily and accurately identify the position coordinates of the light-emitting marker M#.

[0066] In the above-described embodiment, the light-emitting marker M# may be capable of changing the light emission color. The light emission color of the light-emitting marker M# is changed depending on the target area T. For example, in a target area T in which many white structures are installed, the light-emitting marker M# is made to emit light of a color other than white, for example, red light. This makes it possible to easily identify the position coordinates of the light-emitting marker M#.

[0067] Furthermore, in the above-described embodiment, in the correction mode, two light-emitting body markers M# may be arranged at a specific interval in the target region T. The correction coefficient calculation unit 102 specifies the position coordinates of each of the two light-emitting body markers M# in the captured image TG (region correction image) captured in the correction mode. The correction coefficient calculation unit 102 calculates a correction coefficient to correspond to each of the multiple regions divided according to depth, based on the relationship between the positions of the specified light-emitting body markers M#. This makes it possible to calculate a correction coefficient to correspond to each of the multiple regions divided according to depth.

[0068] (Modification 3 of the embodiment) A third modification of the embodiment will be described. In the above embodiment, a frame F, which is a plurality of smoke detection regions, is set in the correction mode, and a zoom ratio is set for each smoke detection region, but this is not limited to the above. In this modification, the zoom ratio is set by a method different from the method for setting the frame F.

[0069] A method for calculating the reduction ratio by the correction coefficient calculation unit 102 in this modification will be described with reference to Fig. 10 (Fig. 10A, Fig. 10B) and Fig. 11 (Fig. 11A, Fig. 11B). Fig. 10 and Fig. 11 are diagrams for explaining the modification 3 of the embodiment.

[0070] 10A shows an image (captured image TG). As shown in the example of this figure, when the y-axis is set along the depth direction of the image, for example, the position coordinate of the light-emitting marker M#2 is the y-coordinate (ya). FIG. 10B is a graph showing the relationship between the position coordinate (y coordinate) of the light-emitting marker M# in the image (captured image TG) and the length (number of pixels) on the image corresponding to the specified length (here, 1 [m]). The correction coefficient calculation unit 102 determines the range in the captured image TG to which the reduction ratio corresponding to each light-emitting marker M# is applied. For example, it is shown that in the area corresponding to the y coordinate (ya) of the light-emitting marker M#2 in FIG. 10A, 100 [pixels] corresponds to 1 m in real space. It is also shown that, when calculating the reduction ratio in area A that is centered on the y coordinate (ya) and within a range of ±L in the y-axis direction, 100 [pixels] corresponds to 1 [m] in real space.

[0071] The correction coefficient calculation unit 102 divides the captured image TG into regions such that the range ratio of the spatial resolution (here, the reduction ratio) is constant. The regions to be divided from the captured image TG may be arbitrarily determined according to the window size used in the calculation when the fire detection unit 101 detects a fire, the size of the threshold for determining the size of something that indicates the possibility of a fire, and the like. The correction coefficient calculation unit 102 divides the region into spaces to which the same reduction ratio can be applied. For example, when applying the same reduction ratio to a space up to twice the actual measurement value of the reduction ratio, the correction coefficient calculation unit 102 applies, for example, a reduction ratio corresponding to 200 [pixels / m] to a region where the length (number of pixels) on the image corresponding to the specified length (here, 1 [m]) is 200 to 100 [pixels / m]. Also, the correction coefficient calculation unit 102 applies, for example, a reduction ratio corresponding to 100 [pixels / m] to a region where the length (number of pixels) on the image corresponding to the specified length (here, 1 [m]) is 100 to 50 [pixels / m]. The correction coefficient calculation unit 102 stores the calculated reduction ratio in the correction coefficient storage unit 103. The correction coefficient storage unit 103 stores the correction coefficients.

[0072] Fig. 11A shows an image (captured image TG). This figure shows area A in the image (captured image TG). Area A corresponds to area A in Fig. 10B, and is an area in the range of ±L in the y-axis direction centered on the y coordinate (ya) of the light-emitting marker M#2.

[0073] FIG. 11B shows an example of information stored in the correction coefficient storage unit 103 of the embodiment. As shown in FIG. 11B, the correction coefficient storage unit 103 stores information corresponding to items such as the y coordinate of the image, the number of pixels (pixel / m) corresponding to 1m, and the reduction ratio (an example of a correction coefficient). The y coordinate of the image indicates a position coordinate in the depth direction in the captured image TG. The number of pixels (pixel / m) corresponding to 1m indicates the number of pixels corresponding to 1m applied to the area indicated by the y coordinate of the image. The reduction ratio (an example of a correction coefficient) indicates the reduction ratio applied to the area indicated by the y coordinate of the image. In the example of this figure, it is shown that a reduction ratio corresponding to 100 [pixel / m] is applied to the area of ​​±L centered on the y coordinate (ya), which corresponds to the area A in FIG. 11A, that is, the area included in the y coordinate (ya-L) to the y coordinate (ya+L).

[0074] In this modification, in the fire detection mode, the fire detection device 10 acquires image information of an imaged image TG in which a target region T is captured. The fire detection device 10 judges whether or not something indicating the possibility of a fire, such as smoke, is captured in the acquired image. When the fire detection device 10 detects an image region including something indicating the possibility of a fire, the fire detection device 10 judges which region of the reduction ratio the lower end of the y coordinate in the image region corresponds to. The fire detection device 10 reduces the image region using the corresponding reduction ratio. Then, the fire detection device 10 calculates the actual size of the thing indicating the possibility of a fire captured in the image. The fire detection device 10 judges whether or not a fire has occurred based on the calculated size of the thing indicating the possibility of a smoke fire. In this way, a fire is detected.

[0075] In the above-described embodiment and its modified example, the case where the image is reduced based on the reduction ratio has been described as an example, but the present invention is not limited to this. The magnification ratio may be calculated, and the calculated magnification ratio may be used to enlarge an area in the depth direction of the image.

[0076] The fire detection system 1 and the fire detection device 10 in the above-mentioned embodiment may be realized in whole or in part by a computer. In that case, a program for realizing the function may be recorded in a computer-readable recording medium, and the program recorded in the recording medium may be read into a computer system and executed to realize the function. The term "computer system" as used herein includes hardware such as an OS and peripheral devices. The term "computer-readable recording medium" refers to portable media such as flexible disks, optical magnetic disks, ROMs, and CD-ROMs, and storage devices such as hard disks built into a computer system. The term "computer-readable recording medium" may also include a medium that dynamically holds a program for a short period of time, such as a communication line when a program is transmitted via a network such as the Internet or a communication line such as a telephone line, and a medium that holds a program for a certain period of time, such as a volatile memory inside a computer system that is a server or client in that case. The above-mentioned program may be a program for realizing a part of the above-mentioned function, or may be a program that can realize the above-mentioned function in combination with a program already recorded in the computer system, or may be a program that is realized using a programmable logic device such as an FPGA.

[0077] Although an embodiment of the present invention has been described in detail above with reference to the drawings, the specific configuration is not limited to this embodiment, and designs that do not deviate from the gist of the present invention are also included. [Explanation of symbols]

[0078] 1. Fire detection system 10. Fire detection equipment 100: Image acquisition unit 101...Fire detection unit 102...Correction coefficient calculation unit (correction setting unit) 103...Correction coefficient storage unit CA…Camera (imaging section)

Claims

1. A fire detection system for detecting a fire using an image of a target area, comprising: An imaging unit that images the target area under a specific brightness condition; a correction setting unit that sets a magnification / reduction ratio for each predetermined area in the captured image captured by the imaging unit; a fire detection unit that detects a fire in the target area by using an image that is enlarged or reduced based on the enlargement / reduction ratio with respect to the captured image captured by the imaging unit under a first brightness condition; Equipped with In a correction mode in which the enlargement / reduction ratio is set by the correction setting unit, The imaging unit images the target area in which a light-emitting marker is arranged under a second brightness condition that is lower in brightness than the first brightness condition, the correction setting unit sets the predetermined area based on position coordinates of the light-emitting marker in an area correction image in which the target area is captured by the imaging unit under the second brightness condition, and sets the enlargement / reduction ratio in the predetermined area calculated based on a size of the light-emitting marker in the area correction image. A fire detection system comprising:

2. In the correction mode, The imaging unit further images the target area under the first brightness condition, the correction setting unit specifies position coordinates of the light-emitting marker using the area correction image, and generates a composite image in which the position coordinates of the light-emitting marker are indicated in the captured image captured by the imaging unit under the first brightness condition. The fire detection system of claim 1 .

3. A fire detection system for detecting a fire using an image of a target area, comprising: An imaging unit that images the target area under a specific brightness condition; a correction setting unit that sets a magnification / reduction ratio for each predetermined area in the captured image captured by the imaging unit; a fire detection unit that detects a fire in the target area by using an image that is enlarged or reduced based on the enlargement / reduction ratio with respect to the captured image captured by the imaging unit under a first brightness condition; Equipped with In a correction mode in which the enlargement / reduction ratio is calculated by the correction setting unit, A light-emitting marker is placed at a specific position in the target area; The imaging unit captures an image of a change in the light emitted by the light-emitting marker in a specific pattern in a time series; the correction setting unit sets the predetermined area based on position coordinates of the light-emitting marker in the area correction image by using a plurality of area correction images captured in time series by the imaging unit, and sets the enlargement / reduction ratio in the predetermined area calculated based on the size of the light-emitting marker in the area correction image. A fire detection system comprising:

4. The imaging unit images the target area under a specific brightness condition, images the target area under a first brightness condition in a fire detection mode in which a fire is detected by the fire detection unit, and images the target area under a second brightness condition that is lower in brightness than the first brightness condition in the correction mode.

4. The fire detection system of claim 3.

5. the correction setting unit extracts candidates for position coordinates of the light-emitting marker using a first area correction image from among a plurality of area correction images captured in time series by the imaging unit, performs frequency analysis based on time-series changes in pixel values ​​of light-emitting pixels corresponding to the extracted candidates for position coordinates, and identifies the position coordinates of the light-emitting marker; 4. The fire detection system of claim 3.

6. The light-emitting marker is capable of changing the light-emitting color. A fire detection system according to any one of claims 1 to 5.

7. The light-emitting marker is composed of two light-emitting bodies spaced apart by a specific distance, the correction setting unit specifies position coordinates of the two separated light-emitting bodies in the area correction image, and calculates the enlargement / reduction ratio based on a relationship between the number of pixels between the position coordinates of the specified light-emitting bodies and the interval; A fire detection system according to any one of claims 1 to 5.

8. A marker position specifying method executed by a computer for use in a fire detection system that detects a fire using an image in which a target area is captured, comprising: An imaging unit captures an image of the target area under a specific brightness condition, a correction setting unit setting a magnification / reduction ratio for each predetermined area in the captured image captured by the imaging unit; a fire detection unit that detects a fire in the target area by using the captured image captured by the imaging unit under a first brightness condition; In a correction mode in which the enlargement / reduction ratio is set by the correction setting unit, The imaging unit images the target area in which a light-emitting marker is arranged under a second brightness condition that is lower in brightness than the first brightness condition, the correction setting unit sets the predetermined area based on position coordinates of the light-emitting marker in an area correction image in which the target area is captured by the imaging unit under the second brightness condition, and sets the enlargement / reduction ratio in the predetermined area calculated based on a size of the light-emitting marker in the area correction image. Marker location methods.

9. A computer used in a fire detection system that detects a fire using an image in which a target area is captured, Imaging the target area under specific lighting conditions; setting a zoom ratio for each predetermined area in the captured image in which the target area is captured; Detecting a fire in the target area by using an image that is enlarged or reduced based on the enlargement / reduction ratio for the captured image captured under a first brightness condition; In the correction mode in which the enlargement / reduction ratio is set or calculated, capturing an image of the target area in which a light-emitting marker is disposed under a second brightness condition that is lower in brightness than the first brightness condition; using an area correction image in which the target area is captured under the second brightness condition, setting the predetermined area based on position coordinates of the light-emitting marker in the area correction image, and setting the enlargement / reduction ratio in the predetermined area calculated based on the size of the light-emitting marker in the area correction image; program.

Citation Information

Patent Citations

  • Smoke detection device and smoke identification method

    JP2020057236A

Cited By

  • Fire detection method and device, electronic equipment and storage medium

    CN122157415A