JP7294848B2 - Monitoring device, monitoring system and monitoring method - Google Patents

Description

The present invention relates to a monitoring device, a monitoring system and a monitoring method.

At oil refineries, surplus gas emitted during the oil refining process is incinerated in a facility called a flare stack. Flames (flares) and black smoke produced by combustion of surplus gas are emitted from the tip of the burner portion of the flare stack. Flare stacks monitor the condition of flares and the generation of black smoke to prevent accidents and protect the environment. In many cases, flares and the like are monitored by photographing the surroundings of the burner with a surveillance camera, and the surveillance personnel confirming the images.

For example, Patent Literature 1 describes a black smoke detection system that detects black smoke emitted from flare stacks and chimneys in chemical plants, thermal power plants, and dust incineration facilities. The black smoke detection system described in US Pat. No. 5,900,000 uses serially captured images of emitted smoke to compute an optical flow estimate for a given region of the two images. Then, the velocity field vector is calculated from the optical flow estimation, and the exhausted smoke area is extracted from the area where the velocity field exists. Further, a luminance histogram is calculated for all pixels in the exhausted smoke area, and when pixels having dark luminance values exist for a certain number or more, it is determined that black smoke exists in the exhausted smoke area.

JP-A-2001-256475

In the above-described flare monitoring, it is required to detect flare abnormalities at an early stage with high accuracy. For example, when trying to detect a flare from an image of a chimney, there is a possibility that the sunset clouds reflected near the chimney may be misdetected as a flare. No technology provided. Also, Japanese Patent Application Laid-Open No. 2002-200002 does not describe a method for detecting an abnormality in flare.

SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide a monitoring device, a monitoring system, and a monitoring method that can solve the above-described problems.

According to one aspect of the present invention, the monitoring device includes an image acquisition unit that acquires an image, and an image analysis unit that analyzes the image and recognizes a predetermined abnormal flare region set with reference to a burner unit that emits flare. an abnormal flare determination unit that detects an abnormal flare based on the size of the flare that occupies the abnormal flare region; and a temporal change in the abnormal flare based on the plurality of images acquired during a predetermined period. and an erroneous detection determination unit that determines the possibility of erroneous detection of the abnormal flare by analyzing the number of pixels corresponding to the flare in the plurality of images per unit time. If the normalized value obtained by dividing the value indicating the variation by the number of pixels is larger than a predetermined threshold value, it is determined that the abnormal flare detection is not an erroneous detection.

According to one aspect of the present invention, the monitoring device includes an image acquisition unit that acquires an image, and an image analysis unit that analyzes the image and recognizes a predetermined abnormal flare region set with reference to a burner unit that emits flare. an abnormal flare determination unit that detects an abnormal flare based on the size of the flare that occupies the abnormal flare region; and a temporal change in the abnormal flare based on the plurality of images acquired during a predetermined period. and an erroneous detection determination unit that determines the possibility of erroneous detection of the abnormal flare by analyzing the number of pixels corresponding to the flare in the plurality of images per unit time. When the normalized value obtained by dividing the value indicating the variation by the number of pixels is equal to or less than a predetermined threshold value, it is determined that there is a possibility of erroneous detection of the abnormal flare.

According to one aspect of the present invention, the monitoring device includes an image acquisition unit that acquires an image, and an image analysis unit that analyzes the image and recognizes a predetermined abnormal flare region set with reference to a burner unit that emits flare. an abnormal flare determination unit that detects an abnormal flare based on the size of the flare that occupies the abnormal flare region; and a temporal change in the abnormal flare based on the plurality of images acquired during a predetermined period. an erroneous detection determination unit configured to determine a possibility of erroneous detection of the abnormal flare by analyzing the erroneous detection determination unit, wherein the erroneous detection determination unit adds the behavior of the flare included in the plurality of images to the images. is included, the possibility of erroneous detection is determined by comparing with the behavior of the cloud .

According to one aspect of the present invention, the monitoring device includes an image acquisition unit that acquires an image, and an image analysis unit that analyzes the image and recognizes a predetermined abnormal flare region set with reference to a burner unit that emits flare. an abnormal flare determination unit that detects an abnormal flare based on the size of the flare that occupies the abnormal flare region; and a temporal change in the abnormal flare based on the plurality of images acquired during a predetermined period. an erroneous detection determination unit that determines a possibility of erroneous detection of the abnormal flare by analyzing the erroneous detection determination unit, wherein the erroneous detection determination unit determines the shape of the flare included in the image in which the abnormal flare is detected. The possibility of erroneous detection is determined based on the degree of left-right symmetry.

According to one aspect of the present invention, the monitoring device includes an image acquisition unit that acquires an image, and an image analysis unit that analyzes the image and recognizes a predetermined abnormal flare region set with reference to a burner unit that emits flare. an abnormal flare determination unit that detects an abnormal flare based on the size of the flare that occupies the abnormal flare region; and a temporal change in the abnormal flare based on the plurality of images acquired during a predetermined period. an erroneous detection determination unit configured to determine a possibility of erroneous detection of the abnormal flare by analyzing, wherein the image analysis unit determines the RGB luminance value corresponding to the flare in the image and the range of the RGB luminance value. based on the setting information indicating, in the abnormal flare region, the number of pixels in the region where the flare is reflected, and the abnormal flare determination unit determines the abnormal flare when the number of pixels is equal to or greater than a predetermined threshold .

According to one aspect of the present invention, the erroneous detection determination unit determines erroneous detection based on the extent to which the hue of the flare included in the video in which the abnormal flare is detected deviates from a predetermined reference hue. Determine feasibility.

According to one aspect of the present invention, the erroneous detection determination unit determines the possibility of erroneous detection based on a trained model constructed by learning the video when the abnormal flare is erroneously detected.

According to one aspect of the present invention, a monitoring system includes an imaging device that captures an image around a flare emitting portion of a flare stack, and any of the monitoring devices described above that acquires the image.

According to one aspect of the present invention, a monitoring method includes the steps of obtaining an image, analyzing the image and recognizing a predetermined abnormal flare region set with reference to a burner section that emits a flare, and detecting the abnormal flare based on the size of the flare occupying the abnormal flare region; and determining the possibility of erroneous detection of flare, wherein the step of determining the possibility of erroneous detection includes a value indicating variation per unit time of the number of pixels corresponding to the flare in the plurality of images. is larger than a predetermined threshold value after normalization by dividing by the number of pixels, it is determined that the abnormal flare detection is not an erroneous detection.

According to one aspect of the present invention, a monitoring method includes the steps of obtaining an image, analyzing the image and recognizing a predetermined abnormal flare region set with reference to a burner section that emits a flare, and detecting the abnormal flare based on the size of the flare occupying the abnormal flare region; and determining the possibility of false detection of flare, wherein the step of determining the possibility of false detection includes determining the behavior of the flare included in the plurality of images, when the images include clouds. to determine the possibility of erroneous detection by comparing with the behavior exhibited by the cloud.

According to one aspect of the present invention, a monitoring method includes the steps of obtaining an image, analyzing the image and recognizing a predetermined abnormal flare region set with reference to a burner section that emits a flare, and detecting the abnormal flare based on the size of the flare occupying the abnormal flare region; and determining the possibility of erroneous detection of flare, wherein the step of determining the possibility of erroneous detection includes determining the degree of left-right symmetry of the shape of the flare included in the image in which the abnormal flare is detected. to determine the possibility of false detection.

According to one aspect of the present invention, a monitoring method includes the steps of obtaining an image, analyzing the image and recognizing a predetermined abnormal flare region set with reference to a burner section that emits a flare, and detecting the abnormal flare based on the size of the flare occupying the abnormal flare region; and determining the possibility of erroneous detection of flare, wherein in the step of recognizing the abnormal flare region, RGB luminance values corresponding to the flare in the video image and setting information indicating the range of the RGB luminance values. Based on this, in the step of calculating the number of pixels in the area in which the flare appears in the abnormal flare area and detecting the abnormal flare, the abnormal flare is determined when the number of pixels is equal to or greater than a predetermined threshold.

According to the present invention, erroneous detection of abnormal flare can be prevented.

It is a figure which shows an example of the monitoring system in one Embodiment of this invention. It is a functional block diagram showing an example of a monitoring device in a first embodiment of the present invention. It is a figure which shows an example of the image|video which the monitoring apparatus in 1st embodiment of this invention analyzes. It is a figure which shows an example of the monitoring screen of the monitoring system in 1st embodiment of this invention. It is the 1st figure explaining the erroneous detection judging processing by the shape in the first embodiment of the present invention. FIG. 10 is a second diagram illustrating erroneous detection determination processing based on shape in the first embodiment of the present invention; FIG. 10 is a first diagram illustrating erroneous detection determination processing based on hue in the first embodiment of the present invention; FIG. 10 is a second diagram illustrating erroneous detection determination processing based on hue in the first embodiment of the present invention; FIG. 11 is a third diagram illustrating erroneous detection determination processing based on hue in the first embodiment of the present invention; It is a figure explaining the false detection determination process by the characteristic of a change with time in 1st embodiment of this invention. It is a flow chart which shows an example of false alarm judgment processing in a first embodiment of the present invention. It is a flowchart which shows an example of the monitoring process in 1st embodiment of this invention. It is a functional block diagram which shows an example of the monitoring apparatus in 2nd embodiment of this invention. It is a figure explaining the monitoring process in 2nd embodiment of this invention.

A flare stack monitoring system according to a first embodiment of the present invention will now be described with reference to FIGS. 1 to 12. FIG.
(System configuration)
FIG. 1 is a diagram showing an example of a monitoring system according to an embodiment of the invention.
FIG. 1 shows the configuration of a monitoring system 1. As shown in FIG. As shown in FIG. 1 , the surveillance system 1 includes a camera 10 , an image acquisition device 15 , an image distribution device 20 , a surveillance monitor 25 and a surveillance device 30 .
The camera 10 captures an image of the burner portion at the tip of the flare stack from which flares and smoke are emitted (flare radiating tip, chimney) and its surroundings. FIG. 3 shows an example of an image captured by the camera 10. As shown in FIG. An image captured by the camera 10 is hereinafter referred to as a flare image. It should be noted that the camera 10 is preferably an imaging device that captures moving images, but it may also be, for example, a device that captures still images at predetermined time intervals.

The image acquisition device 15 acquires the flare image captured by the camera 10 and outputs the flare image to the image distribution device 20 . In general, the image acquisition device 15 is directly connected to the monitor 25 for monitoring, and the monitor looks at the image of the flare displayed on the monitor 25 for monitoring and confirms the vicinity of the burner part of the flare stack and the sky above, and confirms that the flare has occurred. Visually monitor whether it is large, misfired, black smoke, etc. In the monitoring system 1 of this embodiment, the video acquisition device 15 is connected to the video distribution device 20 .

The image distribution device 20 has a function of distributing images. The video distribution device 20 is connected to a monitoring monitor 25 and a monitoring device 30 . The video distribution device 20 outputs the flare video acquired from the video acquisition device 15 to the monitoring monitor 25 and the monitoring device 30 .
The monitoring monitor 25 displays flare images. The surveillance staff watches the flare image displayed on the surveillance monitor 25 .

The monitoring device 30 analyzes the flare image and detects an abnormality in the size of the flare reflected in the flare image, misfire of the flare, and generation of black smoke. When the monitoring device 30 detects a flare abnormality or the like, the monitoring device 30 displays the detection result on a display device connected to the monitoring device 30 and notifies the monitoring staff. Next, the monitoring device 30 will be described in detail.

<First Embodiment>
(Function of monitoring device)
FIG. 2 is a functional block diagram showing an example of a monitoring device according to the first embodiment of the invention.
The monitoring device 30 is a computer such as a PC (Personal Computer) or a server having an arithmetic device such as a CPU (Central Processing Unit). As illustrated, the monitoring device 30 includes an image acquisition unit 31, a setting information acquisition unit 32, an image analysis unit 33, an abnormal flare determination unit 34, an erroneous detection determination unit 341, a misfire determination unit 35, and black smoke. A determination unit 36 , a storage unit 37 and an output control unit 38 are provided.

The image acquisition unit 31 acquires flare images. The video acquisition unit 31 outputs the flare video to the video analysis unit 33 .
The setting information acquisition unit 32 analyzes the flare image and acquires setting information necessary for detecting an abnormality of the flare. The setting information acquisition unit 32 records the acquired setting information in the storage unit 37 .

The video analysis unit 33 analyzes the flare video. Now refer to FIG. FIG. 3 is a diagram showing an example of video analyzed by the monitoring device according to the first embodiment of the present invention. An example of the flare image 50 is shown in FIG. Note that the rectangular frames 51 to 54 are image information added to the flare image 50 by the output control unit 38 rather than the flare image 50 itself. (1) For example, the image analysis unit 33 detects the burner portion 56 appearing in the flare image by pattern recognition based on the shape of the burner portion. The image analysis unit 33 sets a predetermined position of the detected burner portion 56 (for example, the center of the burner portion 56 ) as a reference position in the flare image 50 . For example, the image analysis unit 33 records the coordinate information of the reference position when the lower left end of the flare image 50 is set as the origin in the storage unit 37 .

(2) The video analysis unit 33 also sets predetermined ranges based on the reference position in the flare video 50 as a normal flare region 51, an abnormal flare region 52, a black smoke detection region 53, and a background region .
The normal flare region 51 includes, for example, a portion of the burner portion 56 and defines a range in which the size of the flare 55 can be considered normal.
The abnormal flare region 52 is set including, for example, a region further from the reference position than the normal flare region 51 . The abnormal flare region 52 is set to a position where a portion of the flare 55 may reach when it expands. For example, the abnormal flare region 52 is set on the sky side outside the normal flare region 51 with a width wider than that of the normal flare region 51 . The monitoring device 30 determines that the flare 55 is large (abnormal) when the flare 55 expands and a part thereof reaches the abnormal flare region 52 and occupies a predetermined proportion. The abnormal flare region 52 may be set so as not to include the normal flare region 51 as illustrated, or may be set to include the normal flare region 51 . In the following, unless otherwise specified, the case where the abnormal flare region 52 is set so as not to include the normal flare region 51 will be described as an example.

The black smoke detection area 53 is set to include, for example, a position farther from the reference position than the abnormal flare area 52 is. The black smoke detection area 53 is set for a position where black smoke may reach when black smoke is generated. For example, the black smoke detection area 53 is set on the upper sky side outside the abnormal flare area 52 with a width wider than that of the abnormal flare area 52 . This is because the black smoke may reach a wide range due to the influence of the wind in the sky, so that the black smoke can be reliably detected. When the monitoring device 30 detects the presence of black smoke in the black smoke detection area 53, it determines that black smoke is generated. In order to detect the presence of black smoke, the video analysis unit 33 uses the luminance value of the pixels included in the background area 54 as a reference, and the luminance value of the pixels existing in the black smoke detection area 53 is set to the reference luminance. If it is less than or equal to the predetermined value (if it is darker) than the value, it is determined that the pixel in the black smoke detection area 53 represents black smoke. The black smoke detection area 53 may be set so as not to include the abnormal flare area 52 and the normal flare area 51 as shown in the figure, or the abnormal flare area 52 or the abnormal flare area 52 and the normal flare area 51 may be set. may be set including In the following, unless otherwise specified, the case where the black smoke detection area 53 is set so as not to include the abnormal flare area 52 and the normal flare area 51 will be described as an example.

Background area 54 is set for black smoke detection as described above. A background area 54 is set to an area where flares and black smoke do not reach. The background area 54 is set, for example, for a part of the space that is sufficiently away from the burner section 56 and has no structure.
The normal flare area 51, the abnormal flare area 52, the black smoke detection area 53, and the background area 54 can be arbitrarily set by the observer.

(3) The video analysis unit 33 also detects the flare 55 from the flare video 50 based on the RGB luminance values (values of 256 gradations for each of red, green, and blue in the RGB color space) corresponding to the flare. The RGB brightness values corresponding to the flare and the values defining the range thereof may be input to the monitoring device 30 by the monitor, and the setting information acquisition unit 32 may acquire the values. It is also possible to analyze the pixels in which the flare 55 is reflected among the pixels around the portion 56 and set the RGB luminance values corresponding to the flare.

Based on the analysis result of the image analysis unit 33, the abnormal flare determination unit 34 determines that the flare 55 is an abnormal flare (large flare) when the size of the flare 55 in the abnormal flare region 52 is equal to or greater than a predetermined ratio. More specifically, the abnormal flare determination unit 34 determines the flare 55 as an abnormal flare when the number of pixels having RGB luminance values corresponding to the flare in the abnormal flare region 52 is equal to or greater than a threshold value.

The erroneous detection determination unit 341 determines the possibility of erroneous detection of abnormal flare based on the analysis result of the flare image in which the abnormal flare is detected. For example, the erroneous detection determination unit 341 determines the possibility of erroneous detection based on the shape of the flare (horizontal symmetry) in the flare image. Further, the erroneous detection determination unit 341 determines the possibility of erroneous detection based on the degree of divergence between the hue of the flare reflected in the flare image and the reference hue (45°). Further, the erroneous detection determination unit 341 determines the possibility of erroneous detection based on the characteristics of the temporal change of the flare reflected in the flare image. Processing by the erroneous detection determination unit 341 will be described later with reference to FIGS. 5 to 11. FIG.

Based on the analysis result of the video analysis unit 33, the misfire determination unit 35 determines that the flare 55 has misfired when the size of the flare occupying the normal flare region 51 is equal to or less than a predetermined threshold value. More specifically, the misfire determination unit 35 determines that a misfire has occurred when the number of pixels having RGB luminance values corresponding to flare in the normal flare region 51 is equal to or less than a threshold. Alternatively, the misfire determination unit 35 determines that the number of pixels having RGB luminance values corresponding to the flare is equal to or less than the threshold based on the analysis result of the image analysis unit 33 for the flare image 50 captured by the camera 10. If it continues for a predetermined time or longer, it is determined as a misfire.

Based on the analysis result of the video analysis unit 33, the black smoke determination unit 36 determines that black smoke has occurred when the number of pixels having luminance values corresponding to black smoke in the black smoke detection area 53 is equal to or greater than a threshold. As will be described later, luminance values of 256 gradations between black and white (black and white luminance values) are used to determine black smoke.

The storage unit 37 stores various information. For example, the storage unit 37 stores coordinate information of the reference position of the burner unit 56, setting information of RGB luminance values corresponding to flare and their range, and the like.

The output control unit 38 is a user interface (eg, the 4 monitor screen 60) is generated. The output control unit 38 outputs the user interface image to the display device of the monitoring device 30 for display. FIG. 4 shows an example of the monitoring screen.

(Example of monitoring device output)
FIG. 4 is a diagram showing an example of a monitoring screen of the monitoring system according to the first embodiment of the invention.
The output control unit 38 generates an image of a monitoring screen 60 illustrated in FIG. 4 . The monitoring screen 60 includes a result display field 61, flare luminance display fields 62a-62b, setting information input fields 63a-63f, a registration button 64, and a flare image 50. FIG.
In the result display column 61, the determination results by the abnormal flare determination section 34, the misfire determination section 35, and the black smoke determination section 36 are displayed. When the abnormal flare determination unit 34 does not detect an abnormal flare, the misfire determination unit 35 does not detect misfire, and the black smoke determination unit 36 does not detect black smoke, the output control unit 38 displays character data in the result display field 61. Display "Normal". When the abnormal flare determination unit 34 detects an abnormal flare, the output control unit 38 displays, for example, character data "abnormal flare". Further, when the erroneous detection determination unit 341 detects the possibility of erroneous detection of flare abnormality, the output control unit 38 displays, for example, character data "flare abnormality (possibility of false alarm)". When the misfire determination unit 35 detects a misfire, the output control unit 38 displays, for example, character data "misfire". When the black smoke determination unit 36 detects black smoke, the output control unit 38 displays, for example, character data "black smoke occurs". Further, when the abnormal flare determination unit 34 detects an abnormal flare and the black smoke determination unit 36 detects black smoke, the output control unit 38 displays, for example, character data "abnormal flare and black smoke".

The output control unit 38 displays the RGB luminance values (for example, average values) of the pixels of the portion corresponding to the flare 55 of the flare image 50 in the flare luminance display columns 62a-62b. For example, the output control unit 38 displays the red luminance value in the flare luminance display field 62a and the green luminance value in the flare luminance display field 62b.
The output control unit 38 displays an image in which boundary lines of the normal flare area 51, the abnormal flare area 52, the black smoke detection area 53, and the background area 54 are superimposed on the flare image 50 captured by the camera 10. FIG.
In addition to the flare image 50 displayed on the monitoring monitor 25, the observer watches the result display column 61, the flare image 50, and the flare brightness display columns 62a-62b to monitor the flare. In the flare image 50 displayed on the monitoring monitor 25, the boundary lines indicating the normal flare area 51, the abnormal flare area 52, the black smoke detection area 53, and the background area 54 are not displayed.

The observer sets the threshold for the abnormal flare determination unit 34 to determine that the flare is abnormal, that is, the threshold for the ratio of the number of pixels that can be regarded as showing the flare 55 to the total number of pixels in the abnormal flare region 52. Input in the setting information input field 63a.
The observer determines the threshold at which the black smoke determination unit 36 determines that black smoke has occurred, that is, the ratio of the number of pixels that can be regarded as showing black smoke to the total number of pixels in the black smoke detection area 53. Enter the threshold value in the setting information input field 63b.

The observer inputs the thresholds used when the video analysis unit 33 detects the flare 55 from the flare video 50 into the setting information input fields 63c to 63e. The observer inputs, for example, the red (R) luminance value threshold among the RGB luminance values of the flare 55 in the setting information input field 63c, and the RGB luminance of the flare 55 in the setting information input field 63d. Enter the threshold value for the green (G) luminance value. The observer inputs the brightness range of the threshold values entered in the setting information input fields 63c and 63d into the setting information input field 63e. Based on these inputs, the video analysis unit 33 inputs a red luminance value having an allowable range centered on the value input in the setting information input field 63c and the value input in the setting information input field 63e, and inputs it in the setting information input field 63d. A pixel having a green luminance value whose allowable range is the value entered in the setting information input field 63e with the value entered as the center is recognized as the flare 55. FIG.

The observer inputs a value used when the video analysis unit 33 detects black smoke in the flare video 50 into the setting information input field 63f. The video analysis unit 33 subtracts the value input in the setting information input field 63f from the luminance value after converting the image in the background area 54 into a black and white image by grayscaling the image in the RGB color space (RGB image). Pixels in the black smoke detection area 53 (pixels after grayscaling) having luminance values less than or equal to the value are recognized as pixels in which black smoke is reflected.

When the observer presses the registration button 64, the setting information acquisition unit 32 records the values entered in the setting information input fields 63c to 63f in the storage unit 37. FIG. The video analysis unit 33 analyzes the flare video 50 based on each setting information recorded in the storage unit 37 .

Next, processing for confirming whether or not the detection is appropriate when the abnormal flare determination unit 34 detects an abnormal flare in the analysis of the flare image 50 will be described.
(Determination of false detection)
1. Determination of erroneous detection based on shape FIG. 5 is a first diagram for explaining erroneous detection determination processing based on shape in the first embodiment of the present invention.
An image 5 a is a flare image acquired by the image acquisition unit 31 . Image 5b is an image obtained by displaying pixels having RGB luminance values corresponding to flare in image 5a in white and the rest in black. Images 5a and 5b are examples of abnormal flares. As can be seen from the image 5b, the shape of the abnormal flare is approximately bilaterally symmetrical with the burner section 56 as a reference. Most abnormal flares are observed in a symmetrical shape as illustrated in FIG.

FIG. 6 is a second diagram illustrating erroneous detection determination processing based on shape in the first embodiment of the present invention.
An image 6 a is a flare image acquired by the image acquisition unit 31 . Image 6b is an image in which pixels having RGB luminance values corresponding to the flare are displayed in white with respect to image 6a, and the rest are displayed in black. Images 6a and 6b are examples in which a sunset cloud is erroneously detected as an abnormal flare. As can be seen from the image 6b, the shape of the sunset cloud is often not symmetrical.

As illustrated in FIG. 6, if abnormal flare is detected only by the number of pixels having RGB luminance values in the abnormal flare region 52, confusion may occur with sunset clouds and sunrise clouds photographed with similar RGB luminance values. have a nature. Therefore, the erroneous detection determination unit 341 performs, for example, binarization processing on the flare image in which the abnormal flare is detected by the abnormal flare determination unit 34, and extracts pixels having RGB luminance values within a predetermined range corresponding to the flare. Then, the possibility of erroneous detection of abnormal flare is determined by focusing on the shape formed by the extracted pixels. More specifically, the erroneous detection determination section 341 sets a vertical line to the ground with reference to the proper position of the burner section 56 . The proper position of the burner portion 56 is a position suitable for determining the left-right symmetry of the shape formed by the pixels determined as abnormal flare. For example, the erroneous detection determination section 341 may set the center of the burner section 56 as the appropriate position, or may set the average value of the lateral coordinate information of the leftmost and rightmost pixels among the pixels forming the flare as the appropriate position. . The erroneous detection determination unit 341 divides the flare image into left and right sides of the vertical line to generate images A1 and A2, and mirror-converts one image (for example, image A1) to generate image A1'. Subsequently, the erroneous detection determination unit 341 calculates the degree of similarity between the image A1' and the image A2 by a known method. If the degree of similarity is equal to or greater than a predetermined threshold, the false detection determination unit 341 determines that the images A1 and A2 are bilaterally symmetrical, that is, the abnormal flare detection by the abnormal flare determination unit 34 is not false detection. If the degree of similarity is less than the predetermined threshold, the erroneous detection determination unit 341 determines that the image A1 and the image A2 are not bilaterally symmetrical, that is, that a sunset cloud or the like is detected, and abnormal flare detection is an erroneous detection. determine that it is possible.
Regarding the degree of similarity between the left and right sides, edge processing is performed on the binarized image to extract the outer shape of the image. may be used to determine symmetry if the ratio is within a predetermined range. Alternatively, the horizontal pixel distribution of the detected flare image may be calculated and the kurtosis or skewness thereof may be compared.

In this manner, the erroneous detection determination unit 341 determines the possibility of erroneous detection of abnormal flare based on the shape of the flare extracted from the flare image determined as abnormal flare. This method makes an erroneous judgment when the flare is swept laterally by the wind, or when the shape of the range having the RGB luminance values corresponding to the flare formed by the sunset clouds is nearly symmetrical. there's a possibility that. Therefore, the erroneous detection determination unit 341 may further determine whether or not there is an erroneous detection by a method described below.

2. Judgment of Erroneous Detection Based on Hue FIG. 7 is a first diagram illustrating erroneous detection judgment processing based on hue in the first embodiment of the present invention.
An image 7a is an image of an abnormal flare, and an image 7b is an image when a sunrise cloud is erroneously detected as an abnormal flare. In the case of this example, it is difficult to determine whether or not an erroneous detection has occurred in the determination based on the shape described above. Therefore, the applicant analyzed the hue, saturation, and brightness of pixels determined to be flare in these images. The results are shown in FIG.

FIG. 8 is a second diagram illustrating erroneous detection determination processing based on hue in the first embodiment of the present invention.
The table in FIG. 8 shows the hue, saturation, and brightness of the image 7a in which the flare abnormality is correctly detected, and the hue, saturation, and brightness of the image 7b in which the flare abnormality is erroneously detected. As shown in the figure, it was found that there was no difference in saturation and lightness, but there was a difference in hue. Also, regarding hue, it was found that it is easier to determine whether or not detection is erroneous by focusing on the average value of all pixels rather than the maximum and minimum values among all pixels having RGB luminance values corresponding to flare. In addition to this, the applicant tried a method of discriminating erroneous detection by focusing on the frequency distribution of hues in the monitored area, but no significant difference was obtained between the abnormal flare and the sunrise cloud. Note that the maximum, minimum, and average values of hue in the table of FIG. 8 are values based on a hue of 45°. That is, the average hue value of the flare portion of the image 7a in FIG. 7 is 39.9° (45−5.1), and the average hue value of the flare portion of the image 7b is 45.4°. FIG. 9 shows the hue and saturation to be detected.

FIG. 9 is a third diagram for explaining the erroneous detection determination processing by hue in the first embodiment of the present invention.
As shown in the figure, the saturation of flares, clouds at sunrise, and clouds at sunset is around 25%. Here, a hue of 45° near saturation of 25% corresponds to deep yellow, and the lower the frequency of the hue, the closer to red it becomes. Also, when the saturation is around 25%, the hue of 45° is whitish yellow. As shown in FIG. 8, the sunrise clouds are closer to whitish yellow (hue 45.4°) and the flares are more orange yellow (hue 39.9°).

Therefore, the erroneous detection determining unit 341 extracts all pixels having RGB luminance values within a predetermined range corresponding to the flare from the flare image in which the abnormal flare is detected by the abnormal flare determining unit 34, and determines the hue of all the extracted pixels. to parse Then, the erroneous detection determination unit 341 calculates a corrected hue by subtracting 45° from the analyzed hue, and calculates the average value of the corrected hue after subtraction. When the calculated average value is, for example, ±1°, the false detection determination unit 341 determines that the detected target is a sunrise cloud or a sunset cloud, and that abnormal flare detection is false detection. It may be determined that the case of is not an erroneous detection. Alternatively, if the calculated average value is smaller than a predetermined set value (eg -4), the erroneous detection determination unit 341 determines that the detected flare is an abnormal flare and not an erroneous detection. It may be determined as an erroneous detection.

In addition to determination based on shape, determination based on hue can more accurately prevent erroneous detection of abnormal flare. In addition, since the average value of the hues of all pixels detected as flare is used, it is possible to determine cloud or flare based on whether the image is whitish (cloud) or dark yellow (flare) as a whole. By making the determination based on the average value of the hues of all pixels, it is possible to make a determination similar to the case where a person visually determines whether it is a cloud or a flare.

3. Determination of erroneous detection based on temporal change characteristics The erroneous detection determination unit 341 may also perform erroneous detection determination based on the temporal change of the pixel group detected as abnormal flare.
FIG. 10 is a diagram illustrating erroneous detection determination processing based on characteristics of changes over time in the first embodiment of the present invention.
FIG. 10(a) is a graph showing temporal changes in the ratio of the number of pixels detected as flare to the number of pixels in the entire flare image when clouds at sunset are detected as abnormal flare.
FIG. 10(b) is a graph showing temporal changes in the ratio of the number of pixels detected as flare to the number of pixels in the entire flare image when abnormal flare is correctly detected.
As shown in FIGS. 10A and 10B, the variation in the number of pixels of the sunset cloud is small, and the variation of the number of pixels per unit time of the flare is large. In other words, when observed at small time intervals, flares show drastic changes in behavior, such as fluctuating shape and momentarily disappearing, whereas clouds maintain a constant size and shape, showing changes over time. have characteristics. The false detection determination unit 341 determines the possibility of false alarm by comparing the behavior of flares included in a plurality of images with the behavior of clouds when the images include clouds.

Specifically, the erroneous detection determination unit 341 determines the RGB luminance values corresponding to the flare for the flare video at predetermined intervals (for example, every few seconds to several minutes) when the abnormal flare is detected by the abnormal flare determination unit 34. Calculate the number of pixels it has. Next, the erroneous detection determination unit 341 calculates a temporal change in the calculated number of pixels. If the calculated change in the number of pixels over time is greater than the predetermined range, the erroneous detection determination unit 341 determines that the detected flare is not erroneous detection, and the change in the number of pixels over time is within the predetermined range. If it is within, the erroneous detection determination unit 341 determines that the detected cloud is a sunset cloud or the like and that there is a high possibility of erroneous detection. For example, the erroneous detection determining unit 341 determines that the detected flare is not an erroneous detection if the temporal change in the number of pixels is equal to or greater than the threshold two consecutive times. That is, when the number of pixels at time T1 is P1, the number of pixels at time T2 is P2, and the number of pixels at time T3 is P3, the values of |P1−P2| and |P2−P3| are equal to or greater than the threshold. If so, the erroneous detection determination unit 341 determines that it is not an erroneous detection. It should be noted that the number of consecutive times used as a criterion for determination is not limited to two, and may be arbitrary.
Moreover, it is not necessary to continuously observe that the change in the number of pixels over time exceeds a predetermined range. For example, when the number of times the change in the number of pixels over time exceeds a predetermined range for a predetermined time period (for example, two minutes) is equal to or greater than a threshold value, the false detection determination unit 341 determines that abnormal flare detection is not false detection. good.
Note that the change over time may be calculated as follows. First, the erroneous detection determination unit 341 calculates the standard deviation of the number of flare pixels for a predetermined period of time, and performs normalization processing by dividing the calculated standard deviation by the average value of the number of flare pixels for the predetermined period of time. If the result of the division is equal to or greater than a predetermined threshold value, the erroneous detection determination unit 341 determines that the abnormal flare detection is not an erroneous detection, and if it is equal to or less than the threshold value, it is determined that the abnormal flare detection is an erroneous detection. You may In this normalization process, if the detected object is a cloud, if the cloud is large, the amount of change may be large even if the rate of change is small. This is done in order to accurately judge whether it is a cloud or an abnormal flare by judging by the ratio of the amount to the whole object. In the case of clouds, the ratio of the amount of change is very small, so the value after division is, for example, several percent or less. On the other hand, in the case of an abnormal flare, the amount of change accounts for a large proportion of the entire flare. In the normalization process, the division is performed by the average value of the number of pixels, but the division may be performed using other representative values such as the maximum value, the minimum value, or the median value of the number of pixels in a predetermined time.

Further, the erroneous detection determination unit 341 calculates the ratio of the number of pixels corresponding to the flare to the total number of pixels in the flare image at predetermined time intervals, and determines the abnormal flare determination unit 34 based on the change in the calculated ratio over time. It may be determined whether the abnormal flare detection by is correct or false detection.

In this way, abnormal flares, sunset clouds, and sunrise clouds can be discriminated based on the characteristics of the flare over time indicated by the results of analyzing the video for a predetermined period including the time when the abnormal flare was detected.

Next, an example of false alarm determination processing that combines the above-described three determination methods will be shown.
FIG. 11 is a flowchart showing an example of false alarm determination processing according to the first embodiment of the present invention.
It is assumed that execution/non-execution of the above 1 to 3 erroneous detection determination processes can be set for the erroneous detection determination section 341 . For example, in an environment where the wind is always blowing, the observer may set the camera 10 so that "1. Judgment of erroneous detection based on the shape" is not performed. In the environment where there is little influence of sunrise clouds and sunset clouds due to the installation position, direction, surrounding buildings, etc., only "1. Judgment of false detection based on shape" will be performed and other judgment processing will not be executed. May be set. Of course, the observer may be set to perform only "2. Judgment of erroneous detection based on hue" or only "3. Judgment of erroneous detection based on characteristics of change over time". good. Alternatively, it can be set to execute all the erroneous detection determination processes 1 to 3. These settings are set and registered in a predetermined setting file or the like, and the erroneous detection determination unit 341 performs the following erroneous detection determination processing with reference to the setting contents registered in this setting file. Note that the processing order of 1 to 3 shown below is for convenience, and the erroneous detection determination processing of 1 to 3 can be performed in any order.

The false detection determination section 341 starts false alarm determination processing when the abnormal flare determination section 34 detects an abnormal flare.
First, the erroneous detection determination unit 341 reads the setting file and acquires setting information as to which of the determination processes 1 to 3 is to be executed.
Next, the erroneous detection determination unit 341 determines whether or not to perform determination based on the shape (step S1). If determination is to be made (step S1; Yes), the erroneous detection determination unit 341 performs the processing of "1. Determination of erroneous detection based on shape" (step S2). For example, the erroneous detection determining unit 341 extracts pixels indicating flare from the flare image, and calculates the degree of symmetry (similarity) with respect to the vertical line passing through the burner unit 56 . If it can be regarded as bilaterally symmetrical (similarity is equal to or greater than the threshold), the false detection determination unit 341 determines that it is not false detection. If not symmetrical (similarity is less than the threshold), the erroneous detection determination unit 341 determines that there is a high possibility of erroneous detection. The erroneous detection determination unit 341 associates the method of erroneous detection with the determination result and records them in the storage unit 37 .
If the determination based on the shape is not performed (step S1; No), the process proceeds to step S3.

Next, the erroneous detection determination unit 341 determines whether or not to perform determination based on hue (step S3). If determination based on hue is to be performed (step S3; Yes), the erroneous detection determination unit 341 performs the processing of "2. Determination of erroneous detection based on hue" (step S4). For example, the erroneous detection determination unit 341 extracts pixels indicating flare from the flare image and calculates the average value of the hue. For example, if the average value is around 45° (eg, ±1°), the false detection determination unit 341 has a high possibility of false detection, and the average value is a value smaller than 45° by a predetermined value (eg, 4°) or more. If so, it is determined that it is not an erroneous detection. The erroneous detection determination unit 341 associates the method of erroneous detection with the determination result and records them in the storage unit 37 . On the other hand, if determination based on hue is not performed (step S3; No), the process proceeds to step S5.

Next, the erroneous detection determination unit 341 determines whether or not to perform determination based on the characteristics of changes over time (step S5). If determination based on time change characteristics is to be performed (step S5; Yes), the erroneous detection determination unit 341 performs the process of “3. erroneous detection determination based on time change characteristics” (step S6). For example, the erroneous detection determination unit 341 acquires a flare image captured at a predetermined time after abnormal flare detection, and monitors the transition of the number of pixels having RGB luminance values corresponding to the flare. The erroneous detection determining unit 341 determines that there is no erroneous detection when the variation in the number of pixels exceeds a predetermined range, and that there is a high possibility of erroneous detection when the variation in the number of pixels is seldom seen. The erroneous detection determination unit 341 associates the method of erroneous detection with the determination result and records them in the storage unit 37 . On the other hand, if determination based on the characteristic of change over time is not performed (step S5; No), the process proceeds to step S7.

Next, the erroneous detection determination unit 341 makes a final determination as to whether or not the abnormal flare detection by the abnormal flare determination unit 34 is a false alarm (step S7). The erroneous detection determination unit 341 makes a final determination based on the determination results recorded in the storage unit 37 in steps S2, S4, and S6. For example, when all the methods 1 to 3 are used to determine erroneous detection, and it is determined that there is a possibility of erroneous detection in each of the processes 1 to 3, the erroneous detection determination unit 341 determines that there is a possibility of erroneous detection. Flare detection is determined as a "false alarm". On the other hand, if all the methods 1 to 3 determine that it is not an erroneous detection, the erroneous detection determining unit 341 determines that the abnormal flare detection is not a false alarm.

Further, for example, when the erroneous detection determination is performed by only one of the methods 1 to 3, the erroneous detection determination unit 341 determines that it is a "false alarm" when it is determined that there is a possibility of erroneous detection by that method. , if it is determined not to be an erroneous detection, it is determined to be "not a false alarm". In addition, one method is "1. Judgment of false detection based on shape", and when it is determined that there is a possibility of false detection, while determining that it is not a false alarm, "Possibility of false alarm Supplementary information “Yes” may be added to the determination result.

Also, regarding the judgment results by methods 1 to 3, if it is judged that there is a possibility of erroneous detection in "1. Judgment of erroneous detection based on shape", +1 point, "2. Judgment of erroneous detection based on hue" If it is determined that there is a possibility of false detection by +2 points, and if it is determined that there is a possibility of false detection by “3. , the erroneous detection determining unit 341 may determine that it is a "false alarm" when the total score is equal to or greater than a predetermined number of points, and otherwise determine that it is "not a false alarm".

In addition, when an abnormal flare actually occurs, it is necessary to take countermeasures. If it is determined that one method is not an erroneous detection, even if it is determined as an erroneous detection by the remaining methods, the erroneous detection determination unit 341 emphasizes the abnormal flare detection by the abnormal flare determination unit 34, While judging that it is not a false alarm, it is also possible to add supplementary information to the judgment result that says that there is a possibility of a false alarm.

In addition, in the flowchart of FIG. 11, setting is made as to whether or not to execute each erroneous detection determination process, but a time period may be added to the execution conditions. For example, steps S1 to S7 may be executed only during the time periods of each season when sunrise clouds and sunset clouds are observed (for example, 5:00 to 6:00 and 17:30 to 18:30).

(Flow of monitoring process)
Next, the flow of processing in the monitoring device 30 will be described.
FIG. 12 is a flow chart showing an example of monitoring processing in one embodiment of the present invention. The camera 10 continuously captures an image around the burner portion of the flare stack and outputs the image to the image acquisition device 15 . Also, in the monitoring device 30 , the video acquisition unit 31 acquires the flare video via the video distribution device 20 .

First, the observer sets the positional relationship of the normal flare area 51, the abnormal flare area 52, the black smoke detection area 53, and the background area 54 in the flare image 50 (step S11). For example, the observer sends information defining the positional relationship between the burner portion and the normal flare region 51 (for example, the length of each side of the normal flare region 51, the distance of each side from a predetermined position of the burner portion, etc.) to the monitoring device 30. Enter to 3 and 4 displayed on the display of the monitoring device 30, the monitor directly designates the range of the normal flare region 51 (for example, on the image). , specifying a rectangular frame indicating the normal flare region 51 using an input means such as a mouse, etc.) to set the positional relationship. The same applies to other abnormal flare regions 52, black smoke detection regions 53, and background regions 54. FIG. The setting information acquisition unit 32 acquires these pieces of information and records them in the storage unit 37 .

Next, the video analysis unit 33 analyzes the flare video 50 based on the information that defines the positional relationship between the normal flare region 51, the abnormal flare region 52, the black smoke detection region 53, and the background region 54 recorded in the storage unit 37. are set (step S12). The image analysis unit 33 recognizes the position of the burner portion 56 in the flare image 50 by pattern recognition processing, and sets a predetermined position of the burner portion 56 as a reference point. Specifically, first, the video analysis unit 33 calculates the positional relationship between the reference point and each monitoring area, and calculates the normal flare area 51, abnormal flare area 52, black smoke detection area 53, and background area 54. set. Then, the output control unit 38 displays an image in which the boundary lines of the normal flare area 51, the abnormal flare area 52, the black smoke detection area 53, and the background area 54 are superimposed on the flare image 50 captured by the camera 10 at a predetermined position. An image of the monitoring screen 60 as illustrated in FIG. 4 is generated and displayed on the display.

Next, the monitoring staff inputs each setting information of the monitoring screen 60 illustrated in FIG. The setting information acquisition unit records each input setting information in the storage unit 37 . The video analysis unit 33 sets each setting information recorded in the storage unit 37 as a threshold value used for analyzing the flare video 50 (step S13). Next, the observer starts monitoring (step S14). First, the image analysis unit 33 recognizes the burner portion of the flare image 50, and, if necessary, detects reference points and monitoring areas (normal flare area 51, abnormal flare area 52, black smoke detection area 53, background area 54). is reset (step S15). In this processing, the photographing direction of the camera 10 may be changed by the operation of the observer and then adjusted again to the direction of the flare stack, or the photographing direction of the camera 10 may be deviated due to some external factor. In that case, since each monitoring area was set with respect to the previous reference point, there is a possibility that it does not have an appropriate positional relationship with the current reference point. The processing of step S15 is performed for the purpose of correcting this and adjusting the range of the surveillance area in the flare image 50 captured by the camera 10. FIG.

Next, the video analysis unit 33 analyzes the flare video 50 (step S16). For example, the video analysis unit 33 detects pixels in which the flare 55 is reflected in the normal flare region 51 by the thresholds (for example, red luminance threshold: 243, green luminance threshold: value: 152, brightness width: 30). The video analysis unit 33 outputs the number of pixels in which the flare 55 is reflected in the normal flare region 51 (the number of flare pixels (normal)) to the misfire determination unit 35 .
Further, the video analysis unit 33 detects pixels in which flare appears in the abnormal flare region 52 based on the same threshold value. The video analysis unit 33 outputs the number of pixels in which flare is reflected in the abnormal flare region 52 (the number of flare pixels (abnormal)) to the abnormal flare determination unit 34 .
Although the monitor screen 60 illustrated in FIG. 4 illustrates a case where the red luminance value and the green luminance value are set, the video analysis unit 33 detects pixels in which flare appears in the normal flare region 51 and the abnormal flare region 52. Detection may be performed by adding a blue luminance value to the detection. Also, the monitor screen 60 may be provided with an input field for setting a blue luminance value.

In addition, the video analysis unit 33 grayscales the RGB image for a plurality of pixels in the background area 54 and converts it into a black-and-white image. value, median, etc.). The video analysis unit 33 subtracts the value (for example, 100) input in the setting information input field 63f from the calculated representative value of the black and white luminance values (256 gradations) to determine the pixel value for determining black smoke. Calculate the threshold. The video analysis unit 33 converts all pixels in the black smoke detection area 53 from the RGB image to a black and white image. The video analysis unit 33 compares the black-and-white luminance values of all the pixels in the black smoke detection area 53 in the converted black-and-white image with the black smoke determination threshold. The video analysis unit 33 calculates the number of pixels having luminance values smaller than the black smoke determination threshold and outputs this number of pixels (the number of black smoke pixels) to the black smoke determination unit 36 .

Further, the image analysis unit 33 may process the detection of the flare 55 as follows. That is, the luminance value of the flare 55 is set based on the pixels of the normal flare region 51 instead of using the RGB luminance value thresholds and the luminance width set by the observer. For example, the video analysis unit 33 identifies pixels having a predetermined positional relationship from the reference point (pixels in a portion where the flare is most likely to appear in normal driving conditions) as part of the image in which the flare 55 appears. , the RGB luminance values of the pixel may be calculated and set as the threshold value of the flare 55 . The video analysis unit 33 sets an arbitrary luminance width to the calculated RGB luminance values, and calculates the number of flare pixels (normal) in the normal flare region 51 and the number of flare pixels (abnormal) in the abnormal flare region 52 .

Next, the abnormal flare determination unit 34 determines whether or not the number of flare pixels (abnormal) is equal to or greater than the threshold value for abnormal flare determination (step S17). The threshold value for abnormal flare determination is a value obtained by multiplying the total number of pixels included in the abnormal flare region 52 by the ratio entered in the setting information input field 63a. For example, if the total number of pixels is 100 and 20 (%) is entered in the setting information input field 63a, the abnormal flare determination section 34 determines that the flare pixel number (abnormality) obtained from the video analysis section 33 is 20 or more. determine whether If it is equal to or greater than the threshold value for abnormal flare determination (step S17; Yes), the abnormal flare determination unit 34 outputs the determination result of abnormal flare (the flare 55 is large) to the false detection determination unit 341. The false detection determination unit 341 performs the false alarm determination process described with reference to FIG. 11 (step S171). The false detection determination section 341 outputs the processing result of the false alarm determination processing to the abnormal flare determination section 34 . If the determination result by the erroneous detection determination unit 341 is a false alarm (step S172; Yes), the abnormal flare determination unit 34 determines that the flare is normal, and outputs this determination result to the output control unit 38. Then, the process proceeds to step S19.

If the determination result by the false detection determination unit 341 is not false alarm (step S172; No), the abnormal flare determination unit 34 determines that it is an abnormal flare (step S18). The abnormal flare determination section 34 outputs to the output control section 38 a determination result indicating that an abnormal flare has occurred. If the result of determination by the false detection determination unit 341 is not a false alarm, but additional information indicating that there is a possibility of a false alarm is added, the abnormal flare determination unit 34 determines that an abnormal flare has occurred. Along with the determination result, additional information indicating that there is a possibility of false alarm is output to the output control unit 38 .

If the number of flare pixels (abnormal) included in the abnormal flare region 52 is less than the threshold value for abnormal flare determination (step S17; No), the abnormal flare determination unit 34 determines that the flare is normal, and outputs this determination result. Output to the control unit 38 . Then, the misfire determination unit 35 determines whether or not the number of flare pixels (normal) is equal to or less than the threshold value for misfire determination (step S19). The threshold value for misfire determination is 0, for example. In this case, the misfire determination unit 35 determines whether or not there is a pixel having RGB luminance values that can be regarded as flare in the normal flare region. If the number of flare pixels (normal) is equal to or less than the threshold value for misfire determination (step S19; Yes), the misfire determination unit 35 determines that the flare 55 has misfired (step S20). The misfire determination unit 35 outputs the determination result that the flare 55 has misfired to the output control unit 38 . Note that the threshold value for misfire determination is recorded in the storage unit 37 in advance.

Note that the misfire determination unit 35 makes a misfire determination when the number of flare pixels (normal) acquired from the image analysis unit 33 continues for a predetermined period of time or less than the threshold value for misfire determination, or when the state continues for a predetermined number of times or more. It may be determined that a misfire has occurred when the number of flare pixels (normal) that is equal to or less than the threshold value for is obtained from the video analysis unit 33 .

When the number of flare pixels (normal) included in the normal flare region 51 exceeds the threshold value for misfire determination (step S19; No), the misfire determination unit 35 determines that no misfire has occurred, and controls the output of this determination result. Output to unit 38 . Then, the black smoke determination unit 36 determines whether or not the number of black smoke pixels is equal to or greater than the threshold value for black smoke generation determination (step S21). The threshold value for black smoke generation determination is a value obtained by multiplying the total number of pixels included in the black smoke detection area 53 by the ratio entered in the setting information input field 63b. For example, when the total number of pixels is 100 and 5 (%) is entered in the setting information input field 63b, the black smoke determination unit 36 determines whether the number of black smoke pixels acquired from the video analysis unit 33 is 5 or more. judge. If it is equal to or greater than the threshold value for black smoke generation determination (step S21; Yes), the black smoke determination unit 36 determines that black smoke is generated (step S22). The black smoke determination unit 36 outputs the determination result that black smoke is generated to the output control unit 38 .

When the number of black smoke pixels included in the black smoke detection area 53 is less than the threshold value for black smoke generation determination (step S21; No), the black smoke determination unit 36 determines that black smoke is not generated (step S23 ), and outputs this determination result to the output control unit 38 .

Next, the output control unit 38 reflects the determination result based on each determination process on the monitor screen 60 . For example, if the determinations in steps S17, S19, and S21 are No, the output control unit 38 sets the character data "normal" in the result display column 61 of the monitoring screen 60. FIG. Further, when it is determined that the flare is abnormal in step S18, the output control unit 38 sets, for example, character data "flare abnormal" in the result display field 61 of the monitoring screen 60. FIG. Further, if it is determined in step S171 that there is a possibility of a false alarm although it is not a false alarm, the output control unit 38, for example, displays the text data "flare abnormality (possibility of false alarm)" in the result display column of the monitoring screen 60. Set to 61. Further, when it is determined that there is a misfire in step S<b>20 , the output control unit 38 sets, for example, character data “misfire” in the result display field 61 of the monitoring screen 60 . Further, when it is determined in step S22 that black smoke has occurred, the output control unit 38 sets, for example, character data “black smoke has occurred” in the result display field 61 of the monitoring screen 60 . If abnormal flare is determined in step S18 and black smoke is generated in step S22, the output control unit 38 outputs, for example, character data "flare abnormal and black smoke" or "flare abnormal (possible false alarm)". possible) and occurrence of black smoke” are set in the result display column 61 of the monitoring screen 60.

Further, the output control unit 38 displays the red luminance value of the pixel where the flare 55 is reflected in the normal flare region 51 in the flare luminance display field 62a and the green luminance value in the flare luminance display field 62b. For these values, for example, values calculated by the video analysis unit 33 for the determination in step S19 are used.
The output control unit 38 generates an image of the monitoring screen 60 in which the above values are set, and displays it on the display (step S24). If the monitoring of the flare 55 is to be continued (step S25; No), the processes after step S15 are repeated. It should be noted that when it is determined that "flare abnormality", "misfire", "black smoke" or the like is determined, the time and flare image at that time may be recorded in the storage unit 37 as a log.

Note that when the abnormal flare region 52 is set to include the normal flare region 51, the video analysis unit 33 determines the normal flare region 51 from the number of pixels in the region in which the flare appears in the abnormal flare region 52 in the process of step S16. A value obtained by subtracting the number of pixels in the area where the flare is reflected in , may be output to the abnormal flare determination unit 34 as the number of flare pixels (abnormal). Alternatively, the threshold for abnormal flare determination may be set to a large value.
Further, when the black smoke detection region 53 is set to include the abnormal flare region 52, the video analysis unit 33 determines in step S16 the number of pixels of the black smoke region in the black smoke detection region 53 as an abnormal A value obtained by subtracting the number of pixels in the region where the black smoke is reflected in the flare region 52 may be output to the black smoke determination unit 36 as the number of black smoke pixels. Similarly, when the black smoke detection region 53 is set to include the abnormal flare region 52 and the normal flare region 51, the video analysis unit 33 determines in the process of step S16 that the black smoke in the black smoke detection region 53 is A value obtained by subtracting the number of pixels in the region where black smoke is reflected in the abnormal flare region 52 or the normal flare region 51 from the number of pixels in the region may be output to the black smoke determination unit 36 as the number of black smoke pixels. Alternatively, the threshold for black smoke determination may be set to a large value.

In general, the flare stack shoots images of flares and monitors them. According to the monitoring system 1 of the present embodiment, only by inputting the flare image to the monitoring device 30, it is possible to detect flare abnormalities, misfires, and generation of black smoke with a single system. In addition, as shown in the monitoring system 1 shown in FIG. 1, the flare image is displayed on the monitoring monitor 25 to perform conventional monitoring, while the monitoring device 30 notifies "normal", "flare abnormal", and "flare abnormal". (Possibility of false alarm)”, “Misfire”, “Black smoke”, etc. can be referred to. In addition, the observer selects a normal flare region 51, an abnormal flare region 52, a black smoke detection region 53, and a background region 54, which serve as criteria for judging abnormalities such as flare, depending on the properties of the surplus gas and the installation environment of the flare stack. Since the range and the brightness value corresponding to the flare can be arbitrarily set, it can be applied to any equipment.
It should be noted that once it is determined to be an abnormal flare, then to be determined to be a false alarm, and then to be normal in terms of sampling, the output control unit 38 displays character data such as "normal (flare-similar cloud detected)" on the monitoring screen 60. may be set in the result display field 61 of .

Also, if sunset clouds or sunrise clouds hang behind the flare stack, these clouds may be erroneously detected as abnormal flares. It is possible to prevent confusion between the two and detect an abnormal flare at an early stage with high accuracy. Alternatively, if there is a possibility of false detection, it is possible to notify the observer of the possibility of false detection. You can take action.

<Second embodiment>
A monitoring device 30a according to a second embodiment of the present invention will be described below with reference to FIGS. 13 and 14. FIG.
The monitoring device 30a according to the second embodiment learns the image resulting from the determination of whether or not it is an abnormal flare and whether or not it is an erroneous detection in the processing of the first embodiment, and determines whether it is a normal flare image or an abnormal flare image. It has a function to build a trained model that distinguishes whether an image is an abnormal flare or an image that is falsely detected as an abnormal flare. It also has a function of determining which of the above three types of flare is included in the flare image acquired by the image acquisition unit 31, using the constructed learned image.
Hereinafter, among the configurations according to the second embodiment of the present invention, the same configurations and processes as those of the first embodiment of the present invention are denoted by the same reference numerals, and description thereof will be omitted.

FIG. 13 is a functional block diagram showing an example of a monitoring device according to the second embodiment of the invention.
A monitoring device 30a according to the second embodiment includes an abnormal flare determination section 34a in place of the abnormal flare determination section 34. FIG. In addition to the function of the abnormal flare determination unit 34 of the first embodiment, the abnormal flare determination unit 34a has a function of determining normal flare and abnormal flare based on the learned model. The monitoring device 30a also includes a learning section 39 . These functions will now be described with reference to FIG.

FIG. 14 is a diagram illustrating monitoring processing in the second embodiment of the present invention.
First, the function of the learning section 39 will be described with reference to the upper diagram of FIG. 12, the learning unit 39 assigns “normal”, “abnormal”, “false alarm” or “similar” to the image determined as “normal”, “abnormal flare”, or “false alarm”, respectively. Learning data labeled as . Note that this label may be attached after the observer actually confirms "normal", "abnormal flare", and "false alarm" after the processing of FIG. 12 . Next, the observer instructs the monitoring device 30a to perform learning. Then, the learning unit 39 reads the learning data accumulated in the storage unit 37, performs deep learning, and learns a trained model (for example, multi-layered neural network).

Next, the function of the abnormal flare determination section 34a will be described with reference to the lower diagram in FIG. The abnormal flare determination unit 34a inputs the flare image to be monitored into the trained model and determines "normal", "abnormal", and "false alarm". The abnormal flare determination unit 34a may perform abnormal flare detection processing based on the determination result of the learned model instead of the processing of step S17 in FIG. For example, if the discrimination result of the trained model is "normal" or "false alarm", the processing from step S19 onward is performed in the same manner as when "NO" is determined in step S17. Further, when the discrimination result of the learned model is "abnormal", the processing after step S18 is performed.
Alternatively, if the discrimination result of the learned model is "normal", the process proceeds to step S19 and thereafter, and if the discrimination result of the learned model is "abnormal" or "false alarm", the process from step S171 is performed. can be
Alternatively, the determination by the trained model and each determination process described in the flowchart of FIG. 12 may be performed together. For example, the abnormal flare determination unit 34a performs determination using a learned model while performing determination in step S17. Then, if the determination in step S17 is YES, or if the learned model determines "abnormality" or "false alarm", then processing from step S171 onward may be performed.
Further, while performing such processing, the learning data may be further accumulated, and the learning unit 39 may reconstruct the learned model at a predetermined timing. This makes it possible to improve the accuracy of the learned model.

According to this embodiment, an abnormal flare and an image similar to an abnormal flare can be discriminated by a trained model. As a result, erroneous detection of abnormal flare can be prevented, and abnormal flare can be detected with high accuracy.
It is also possible to build a trained model that can determine misfires and black smoke by learning images of misfires and black smoke.

Each process in the monitoring device 30 can be realized by executing a program by the CPU or the like of the monitoring device 30, for example. The program executed by the monitoring device 30 may be recorded on a computer-readable recording medium, and may be realized by reading and executing the program recorded on this recording medium. Note that the monitoring device 30 includes hardware such as an OS and peripheral devices. The computer-readable recording medium is, for example, a portable medium such as a flexible disk, a magneto-optical disk, a ROM, or a CD-ROM, or a storage device such as a hard disk incorporated in the monitoring device 30 . In addition, a computer-readable recording medium dynamically holds a program for a short period of time, like a communication line for transmitting a program via a network such as the Internet or a communication line such as a telephone line. It may also include something that holds the program for a certain period of time, such as a volatile memory inside a computer system that serves as a server or client in that case. Further, the program may be for realizing part of the functions described above, or may be capable of realizing the functions described above in combination with a program already recorded in the computer system.

In addition, it is possible to appropriately replace the components in the above-described embodiments with well-known components without departing from the scope of the present invention. Moreover, the technical scope of the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the scope of the present invention.
Note that the RGB luminance value is a value represented by a combination of red (R), green (G), and blue (B) luminance values (0 to 255) in the RGB color space. is a value represented by 256 gradations from 0 (black) to 255 (white).

DESCRIPTION OF SYMBOLS 1... Monitoring system 10... Camera 15... Video acquisition apparatus 20... Video distribution apparatus 25... Monitoring monitor 30... Monitoring apparatus 31... Video acquisition part 32... Setting Information acquisition unit 33 Video analysis unit 34, 34a Abnormal flare determination unit 341, 341a False detection determination unit 35 Misfire determination unit 36 Black smoke determination unit 37 Storage Unit 38 Output control unit 39 Learning unit 50 Flare image 51 Normal flare area 52 Abnormal flare area 53 Black smoke detection area 54 Background area 60. Monitoring screen 61 Result display field 62a, 62b Flare luminance display field 63a to 63f Setting information input field 64 Registration button

Claims (12)

an image acquisition unit that acquires an image;
an image analysis unit that analyzes the image and recognizes a predetermined abnormal flare area set with reference to the burner unit that emits the flare;
an abnormal flare determination unit that detects an abnormal flare based on the size of the flare that occupies the abnormal flare region;
an erroneous detection determination unit that determines the possibility of erroneous detection of the abnormal flare by analyzing temporal changes in the abnormal flare based on the plurality of images acquired during a predetermined period;
with
The erroneous detection determination unit divides a value indicating variation per unit time in the number of pixels corresponding to the flare in the plurality of images by the number of pixels after normalization, and if the normalized value is larger than a predetermined threshold, Abnormal flare detection is judged not to be false detection,
surveillance equipment. an image acquisition unit that acquires an image;
an image analysis unit that analyzes the image and recognizes a predetermined abnormal flare area set with reference to the burner unit that emits the flare;
an abnormal flare determination unit that detects an abnormal flare based on the size of the flare that occupies the abnormal flare region;
an erroneous detection determination unit that determines the possibility of erroneous detection of the abnormal flare by analyzing temporal changes in the abnormal flare based on the plurality of images acquired during a predetermined period;
with
The erroneous detection determination unit determines, if a normalized value obtained by dividing a value indicating variation per unit time in the number of pixels corresponding to the flare in the plurality of images by the number of pixels is equal to or less than a predetermined threshold, Abnormal flare detection is judged to have the possibility of false detection,
surveillance equipment. an image acquisition unit that acquires an image;
an image analysis unit that analyzes the image and recognizes a predetermined abnormal flare area set with reference to the burner unit that emits the flare;
an abnormal flare determination unit that detects an abnormal flare based on the size of the flare that occupies the abnormal flare region;
an erroneous detection determination unit that determines the possibility of erroneous detection of the abnormal flare by analyzing temporal changes in the abnormal flare based on the plurality of images acquired during a predetermined period;
with
The false detection determination unit compares the behavior of the flare included in the plurality of images with the behavior of the cloud when the image includes the cloud, and determines the possibility of false detection.
surveillance equipment. an image acquisition unit that acquires an image;
an image analysis unit that analyzes the image and recognizes a predetermined abnormal flare area set with reference to the burner unit that emits the flare;
an abnormal flare determination unit that detects an abnormal flare based on the size of the flare that occupies the abnormal flare region;
an erroneous detection determination unit that determines the possibility of erroneous detection of the abnormal flare by analyzing temporal changes in the abnormal flare based on the plurality of images acquired during a predetermined period;
with
The erroneous detection determination unit determines the possibility of erroneous detection based on the degree of left-right symmetry of the shape of the flare included in the image in which the abnormal flare is detected.
surveillance equipment. an image acquisition unit that acquires an image;
an image analysis unit that analyzes the image and recognizes a predetermined abnormal flare area set with reference to the burner unit that emits the flare;
an abnormal flare determination unit that detects an abnormal flare based on the size of the flare that occupies the abnormal flare region;
an erroneous detection determination unit that determines the possibility of erroneous detection of the abnormal flare by analyzing temporal changes in the abnormal flare based on the plurality of images acquired during a predetermined period;
with
The video analysis unit calculates the number of pixels in the region in which the flare appears in the abnormal flare region based on setting information indicating RGB luminance values corresponding to the flare in the video and the range of the RGB luminance values. ,
The abnormal flare determination unit determines the abnormal flare when the number of pixels exceeds a predetermined threshold.
surveillance equipment. The erroneous detection determination unit determines the possibility of erroneous detection based on the extent to which the hue of the flare included in the image in which the abnormal flare is detected deviates from a predetermined reference hue.
A monitoring device according to any one of claims 1 to 5 . The erroneous detection determination unit determines the possibility of erroneous detection based on a trained model constructed by learning the video when the abnormal flare is erroneously detected.
A monitoring device according to any one of claims 1 to 6 . an imaging device for capturing an image around the flare emitting portion of the flare stack;
a monitoring device according to any one of claims 1 to 7 , which acquires the video;
surveillance system. obtaining an image;
a step of analyzing the image and recognizing a predetermined abnormal flare area set with reference to the burner portion that emits the flare;
detecting an abnormal flare based on the size of the flare occupying the abnormal flare region;
determining the possibility of erroneous detection of the abnormal flare by analyzing the temporal change of the abnormal flare based on the plurality of images acquired during a predetermined period;
has
In the step of determining the possibility of erroneous detection, a standardized value obtained by dividing a value indicating a change per unit time in the number of pixels corresponding to the flare in the plurality of images by the number of pixels is a predetermined threshold value. If it is larger, determining that the abnormal flare detection is not a false detection;
Monitoring method. obtaining an image;
a step of analyzing the image and recognizing a predetermined abnormal flare area set with reference to the burner portion that emits the flare;
detecting an abnormal flare based on the size of the flare occupying the abnormal flare region;
determining the possibility of erroneous detection of the abnormal flare by analyzing the temporal change of the abnormal flare based on the plurality of images acquired during a predetermined period;
has
In the step of determining the possibility of erroneous detection, the behavior of the flare included in the plurality of images is compared with the behavior of clouds when the images include clouds to determine the possibility of erroneous detection. do,
Monitoring method. obtaining an image;
a step of analyzing the image and recognizing a predetermined abnormal flare area set with reference to the burner portion that emits the flare;
detecting an abnormal flare based on the size of the flare occupying the abnormal flare region;
determining the possibility of erroneous detection of the abnormal flare by analyzing the temporal change of the abnormal flare based on the plurality of images acquired during a predetermined period;
has
In the step of determining the possibility of erroneous detection, the possibility of erroneous detection is determined based on the degree of symmetry of the shape of the flare included in the image in which the abnormal flare is detected.
Monitoring method. obtaining an image;
a step of analyzing the image and recognizing a predetermined abnormal flare area set with reference to the burner portion that emits the flare;
detecting an abnormal flare based on the size of the flare occupying the abnormal flare region;
determining the possibility of erroneous detection of the abnormal flare by analyzing the temporal change of the abnormal flare based on the plurality of images acquired during a predetermined period;
has
In the step of recognizing the abnormal flare region, pixels in the region where the flare appears in the abnormal flare region based on setting information indicating RGB luminance values corresponding to the flare in the image and a range of the RGB luminance values. calculate the number,
In the step of detecting the abnormal flare, the abnormal flare is determined when the number of pixels exceeds a predetermined threshold.
Monitoring method.

Images