JP7383531B2 - 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, excess gas emitted during the oil refining process is incinerated in equipment called a flare stack. From the tip of the burner section of the flare stack, flame (flare) and black smoke are emitted from the combustion of excess gas. The flare gas line is connected to piping from multiple units in the refinery. When the seal of the safety valve of the piping deteriorates, gas originally used in the refinery (hereinafter referred to as flare gas) may flow into the flare gas line and be released as flare from the tip of the burner section. This is not preferable from the viewpoint of energy saving. When flare gas is released, the size of the flare becomes larger than normal. Taking advantage of this property, the state of the flare is often photographed by a surveillance camera, and a surveillance officer visually checks the footage to monitor whether flare gas is flowing into the flare gas line.

Patent Document 1 describes a black smoke detection system that detects black smoke discharged from flare stacks and chimneys in chemical plants, thermal power generation facilities, and dust incineration facilities. The black smoke detection system described in Patent Document 1 calculates optical flow estimation for a predetermined region of two images using consecutive images of emitted smoke. Then, a velocity field vector is calculated from the optical flow estimation, and an exhaust smoke region is extracted from the region where the velocity field exists. Further, a brightness histogram is calculated for all pixels in the exhaust smoke area, and if there are a certain number of pixels having dark brightness values, it is determined that black smoke exists in the exhaust smoke area.

Japanese Patent Application Publication No. 2001-256475

In a flare stack, it is necessary to understand the flow rate (loss amount) of flare gas released. The flow rate of flare gas discharged from the flare stack can be monitored by installing a gas flow meter in the flare gas piping. However, installing a new gas flow meter in the flare gas piping must be done during a regular repair period for the equipment, and there is a limit to the period during which the work can be done. Furthermore, the cost required to install a gas flow meter is high. On the other hand, with the method of photographing flares with a surveillance camera and visually checking them by surveillance personnel, it is possible to estimate that a large amount of flare gas is being released based on the size of the flare, but it is not possible to quantitatively understand the flow rate of flare gas. It is not possible. There is a need for a simple and low-cost method to measure the flow rate of flare gas. Patent Document 1 also does not describe a method for measuring the flare gas flow rate.

Therefore, an object of the present invention is to provide a monitoring device, a monitoring system, and a monitoring method that can solve the above-mentioned problems.

According to one aspect of the present invention, a monitoring device includes an image acquisition unit that acquires an image of a flare, a luminance value of a predetermined flare pixel in which a flare is captured, and a luminance value of a predetermined background pixel in which an object other than the flare is captured. Each pixel included in the video is classified into either a first pixel close to the flare pixel or a second pixel close to the background pixel based on the value, and the total number of pixels classified as the first pixel is determined. and a flare gas flow rate estimating unit that estimates the flow rate of flare gas, which is the gas flowing through the flare gas piping , based on the total number of pixels, and compares the flow rate of the flare gas with a predetermined threshold , an output unit that issues an alarm if the flow rate of the flare gas exceeds the threshold ; the image acquisition unit acquires the images taken from two different directions; and the image analysis unit Based on the video shot from the first direction among the video images taken from the first direction, the number of pixels in the horizontal direction W1 and the number of pixels in the vertical direction H1 of the pixel group classified as the first pixel are calculated. Based on the images taken from two directions, the number of pixels in the horizontal direction W2 and the number of pixels in the vertical direction H2 of the pixel group classified as the first pixel are calculated, and the number of pixels H1 and the number of pixels are calculated. The volume V is calculated by multiplying the larger number of pixels in H2, the number of pixels W1, and the number of pixels W2, and the flare gas flow rate estimator calculates the total number of pixels by setting the volume V to the total number of pixels. The flow rate of the flare gas is estimated based on.

According to one aspect of the present invention, a monitoring system includes two imaging devices that capture images of the vicinity of the flare emission part of the flare stack from different directions, and the above-mentioned monitoring device.

According to one aspect of the present invention, a monitoring method includes the steps of acquiring an image of a flare, the brightness value of a predetermined flare pixel in which a flare is captured, and the brightness value of a predetermined background pixel in which a non-flare is captured. Based on the above, each pixel included in the video is classified as either a first pixel close to the flare pixel or a second pixel close to the background pixel, and the total number of pixels classified as the first pixel is calculated. estimating the flow rate of flare gas, which is the gas flowing through the flare gas piping , based on the total number of pixels; and comparing the flow rate of the flare gas with a predetermined threshold value , and determining whether the flow rate of the flare gas exceeds the threshold value. If it exceeds the above , an alarm is issued, and in the acquiring step, the images taken from two different directions are acquired, and in the calculating step, the first one of the acquired images is Based on the image taken from one direction, the number of pixels in the horizontal direction W1 and the number of pixels in the vertical direction H1 of the pixel group classified as the first pixel are calculated, and the number of pixels in the vertical direction H1 is calculated based on the image taken from the second direction. Based on the video, the number of pixels in the horizontal direction W2 and the number of pixels in the vertical direction H2 of the pixel group classified as the first pixel are calculated, and the number of pixels that is larger between the number of pixels H1 and the number of pixels H2 is calculated. , the volume V is calculated by multiplying the number of pixels W1 and the number of pixels W2, and in the estimating step, the volume V is the total number of pixels, and the flow rate of the flare gas is calculated based on the total number of pixels. presume .

According to the present invention, the flow rate of flare gas can be estimated.

FIG. 1 is a diagram illustrating an example of a monitoring system in an embodiment. It is a functional block diagram showing an example of a monitoring device in one embodiment. FIG. 3 is a diagram illustrating an example of a video analyzed by a monitoring device according to an embodiment. FIG. 2 is a first diagram illustrating flare detection processing in an embodiment. FIG. 7 is a second diagram illustrating flare detection processing in an embodiment. FIG. 7 is a third diagram illustrating flare detection processing in an embodiment. FIG. 4 is a fourth diagram illustrating flare detection processing in one embodiment. FIG. 5 is a fifth diagram illustrating flare detection processing in an embodiment. FIG. 3 is a diagram illustrating an example of a measurement result of the number of flare pixels in an embodiment. FIG. 3 is a diagram illustrating processing for estimating a flare gas flow rate in one embodiment. It is a figure showing an example of setting processing in one embodiment. FIG. 2 is a first diagram illustrating an example of flare gas flow rate estimation processing in an embodiment. FIG. 7 is a second diagram illustrating an example of flare gas flow rate estimation processing in one embodiment.

<Embodiment>
Hereinafter, a flare stack monitoring system according to one embodiment will be described with reference to FIGS. 1 to 13.
FIG. 1 is a diagram showing an example of a monitoring system according to an embodiment of the present invention.
FIG. 1 shows the configuration of a monitoring system 1. As shown in FIG. 1, the monitoring system 1 includes a camera 10A, a camera 10B, a video acquisition device 15, a video distribution device 20, a monitoring monitor 25, a monitoring device 30, and a central monitoring device 40. .
As shown in the figure, a camera 10A and a video acquisition device 15, a video acquisition device 15 and a video distribution device 20, a video distribution device 20 and a monitoring monitor 25, a video distribution device 20 and a monitoring device 30, a camera 10B and a monitoring device 30, and a monitoring device. The device 30 and the central monitoring device 40 are connected to each other.

The cameras 10A and 10B take images of the burner section at the tip of the flare stack (the tip section that dissipates the flare, the chimney) and its surroundings. Camera 10A and camera 10B take images around the burner section from different directions. The reason for photographing from different directions is to be able to photograph the three-dimensional appearance of the flare even when the flare is waving in the wind. The images taken by the cameras 10A and 10B are hereinafter referred to as flare images.

Note that the cameras 10A and 10B are preferably imaging devices that take moving images, but may also take still images at predetermined time intervals, for example. Further, either or both of the cameras 10A and 10B may be an infrared camera. In flare stacks, steam is sometimes injected from the burner section to suppress the spread of flare. Injecting steam reduces the flame of the flare, but releases water vapor instead. An infrared camera can detect water vapor and more accurately estimate the flow rate of flare gas. Furthermore, an infrared camera can detect hydrogen gas even when it is released. Although it is desirable to have two cameras, it is also possible to use one of the cameras 10A and 10B.

The video acquisition device 15 acquires the flare video captured by the camera 10A, and outputs the flare video to the video distribution device 20. In general, the image acquisition device 15 is directly connected to the monitoring monitor 25, and the monitoring personnel check the flare image displayed on the monitoring monitor 25 to check the area around and above the burner part of the flare stack, and to detect flares. Visually monitor to see if it is getting bigger, if there is a misfire, if black smoke is being generated, etc. In the monitoring system 1 of this embodiment, the video acquisition device 15 is connected to the video distribution device 20.

The video distribution device 20 has a function of distributing video. 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 observer monitors the flare video displayed on the monitoring monitor 25.

Camera 10B is connected to monitoring device 30 via wireless communication. The camera 10B transmits the captured flare video to the monitoring device 30.

The monitoring device 30 analyzes the flare images taken by the cameras 10A and 10B and estimates the flow rate of the flare gas. When the monitoring device 30 estimates the flow rate of flare gas, it displays the estimation result, an alarm, etc. on the display unit of the monitoring device 30 or a display device connected to the monitoring device 30, and notifies the monitoring personnel. Further, the monitoring device 30 transmits an estimated value of the flow rate of flare gas, an alarm, etc. to the central monitoring device 40. Next, the monitoring device 30 will be explained in detail.

FIG. 2 is a functional block diagram showing an example of a monitoring device according to an embodiment of the present invention.
The monitoring device 30 is a computer such as a PC (Personal Computer) or a server equipped with an arithmetic unit such as a CPU (Central Processing Unit). As illustrated, the monitoring device 30 includes a video acquisition section 31, a setting information acquisition section 32, a video analysis section 33, a flare gas flow rate estimation section 34, a storage section 35, and an output section 36.

The image acquisition unit 31 acquires flare images captured by the cameras 10A and 10B. The video acquisition unit 31 outputs the flare video to the video analysis unit 33.
The setting information acquisition unit 32 acquires various types of setting information necessary for flare video analysis. The setting information acquisition unit 32 records the acquired setting information in the storage unit 35.

The video analysis unit 33 analyzes the flare video. Reference is now made to FIG. FIG. 3 is a diagram showing an example of a video analyzed by a monitoring device according to an embodiment of the present invention. FIG. 3 shows an example of a flare image 50 taken by the camera 10A. Note that the rectangular frames indicated by 52 to 53 do not indicate the flare image 50 itself, but indicate areas necessary for analyzing the flare image. These areas are arranged based on the setting information acquired by the setting information acquisition unit 32.

(1) For example, the video analysis unit 33 detects the burner section 51 appearing in the flare video by pattern recognition based on the shape of the burner section. The video analysis unit 33 sets the detected predetermined position of the burner unit 51 (for example, the center of the tip of the burner unit 51) as a reference position in the flare video 50. The video analysis unit 33 records, in the storage unit 35, coordinate information of a reference position when the lower left end of the flare video 50 is set as the origin, for example. Alternatively, instead of the video analysis unit 33 analyzing the image, a monitoring person may input the coordinate information of the center of the tip of the burner unit 51 to the monitoring device 30.

(2) The video analysis unit 33 sets a predetermined range based on the reference position as the flare monitoring area 52 and the background area 53 based on the setting information acquired by the setting information acquisition unit 32.
The flare monitoring area 52 defines, for example, a range in which the flare 54 is observed from the tip of the burner section 51.
The background area 53 is set to an area where flare and black smoke do not reach. The background area 53 is set, for example, in a part of the space that is sufficiently far away from the burner section 51 and in which no structure exists. The flare monitoring area 52 and the background area 53 can be set arbitrarily by the observer.

(3) The video analysis unit 33 detects the flare 54 from the flare video 50 based on the RGB luminance values (values of 256 gradations of red, green, and blue in the RGB color space) corresponding to the flare. The video analysis unit 33 may analyze the pixels included in the flare monitoring area 52 and set the RGB brightness values corresponding to the flare by processing described later, or the video analysis unit 33 may set the RGB brightness values corresponding to the flare by processing described later A person may input RGB luminance values corresponding to flare into the monitoring device 30, and the setting information acquisition unit 32 may acquire the values.

The flare gas flow rate estimation unit 34 estimates the flare gas flow rate based on the number of pixels of the flare 54 detected in the flare image. Further, the flare gas flow rate estimation unit 34 compares the estimated value of the flare gas flow rate with a plurality of preset threshold values, and instructs the output unit 36 to issue an alarm if the flare gas flow rate exceeds each threshold value. do.

The storage unit 35 stores various information. For example, the storage unit 35 stores coordinate information of the reference position of the burner unit 51, setting information of RGB brightness values corresponding to flare, a calibration curve showing the relationship between the flare gas flow rate and the number of pixels, and the like.

The output unit 36 generates an image of a user interface for displaying the estimation result of the flare gas flow rate by the flare gas flow rate estimating unit 34 and inputting setting information. Further, the output unit 36 outputs an alarm based on an instruction from the flare gas flow rate estimation unit 34.

4 to 8 are first to fifth diagrams each illustrating flare detection processing in one embodiment.
An example of the flare image 50 is shown in FIG. 4(a). FIG. 4B shows an image (hereinafter referred to as a flare image) obtained by analyzing the flare video 50 and extracting only the flare. A method for creating the flare image shown in FIG. 4(b) will be described below. First, the observer sets a range of RGB luminance values corresponding to flare, as shown in FIG. In FIG. 5, R5 indicates the R (red) luminance range, G5 indicates the G (green) luminance range, and B5 indicates the B (blue) luminance range. Next, through image processing, pixels having luminance values included in the set range are extracted as pixels indicating flare. An image is generated by assigning different colors to pixels indicating flare and pixels indicating non-flare. As a result, the flare image shown in FIG. 4(b) is obtained. In the flare image of FIG. 4(b), an annular region is displayed as a flare. As can be seen by comparing FIGS. 4(a) and 4(b), only a portion of the flare can be detected in the flare image of FIG. 4(b).

In this embodiment, in order to improve flare detection accuracy, a three-dimensional space is prepared in which the coordinate axes are the luminance values of R (red), G (green), and B (blue) colors of RGB luminance. Then, in this three-dimensional space, the RGB luminance values of typical pixels in which flare is captured (flare pixels) are plotted, and the RGB luminance values of non-flare pixels in which flare is not captured (background pixels) are plotted, The RGB luminance values of each pixel of the flare image 50 are plotted in this three-dimensional space. Reference is now made to FIG. A point 61 in FIG. 6 is a point where the RGB luminance values of a typical pixel in which flare is captured are plotted. A point 63 is a point where the RGB luminance values of a pixel representing a background such as the sky are plotted. A point 62 is a point where the RGB luminance values of a certain pixel in the flare image 50 are plotted. The video analysis unit 33 calculates the points 62 and 61 based on the coordinates of the point 62 where the RGB brightness values of pixels in the flare video 50 are plotted and the coordinates of the point 61 where the RGB brightness values of typical flares are plotted. The distance between them (the length of the vector 64) is calculated. Further, the video analysis unit 33 determines the distance between the points 62 and 63 based on the coordinates of the point 62 where the RGB brightness values of pixels in the flare video 50 are plotted and the coordinates of the point 63 where the RGB brightness values of the background are plotted. The distance (length of vector 65) is calculated. The video analysis unit 33 then classifies the pixel corresponding to the point 62 into the shorter of the length of the vector 64 and the length of the vector 65. In other words, if the length of the vector 64 is short, the video analysis unit 33 determines that the pixel corresponding to the point 62 is a pixel indicating flare (the first pixel), and if the length of the vector 65 is short, it determines that the pixel corresponding to the point 62 is a non-flare pixel. It is determined that the pixel indicates flare (second pixel). The video analysis unit 33 performs the same determination for all pixels within the flare video 50. The video analysis unit 33 creates a flare image by assigning different colors to pixels determined to be flare and pixels determined to be non-flare. A flare image obtained by such processing is shown in FIG. 7(a).

Referring to the flare image in FIG. 7(a), it can be seen that more areas are detected as flares than in the flare image in FIG. 4(b), and the detected flare area has been improved. . However, in the flare image of FIG. 7(a), a cavity is created at the center of the flare. Next, a process for filling in this cavity will be explained. First, the video analysis unit 33 divides the flare image shown in FIG. 7(a), for example, into widths each having a predetermined number of pixels from top to bottom in the Y-axis direction, and divides the flare image into a row of parts. The following processing is performed on each image. The video analysis unit 33 scans a row of partial images from left to right in the X-axis direction. At that time, pixels indicating flare are detected consecutively for a predetermined number of pixels or more, then pixels indicating non-flare are detected for a predetermined number or more consecutively, and then pixels indicating flare are detected for a predetermined number of pixels or more. If more than the number of pixels are detected consecutively, the video analysis unit 33 determines that the partial image includes a cavity. Then, the pixel group indicating non-flare sandwiched between the pixel group indicating flare is recognized as flare, and the color corresponding to flare (the same color as in Fig. 7(a) or a different color) is assigned to these pixel groups. Good.) is attached. FIG. 7(b) shows a flare image obtained as a result of processing to complement such cavities. Note that H1 and W1 in FIG. 7(b) will be described later.

FIG. 8(a) shows an example of a flare image 50 different from that shown in FIG. 4(a). In this example, a relatively small flare is emitted from the tip of the burner section 51. A method of detecting pixels in which flare is captured based on the distance between each pixel of the flare image 50 and a point indicating a typical flare RGB luminance value and a point indicating a background RGB luminance value, as explained using FIG. 6. If so, it was confirmed that the flare could be detected even when the flare was small, as shown in the flare image of FIG. 8(b).

FIG. 9 is a diagram illustrating an example of measurement results of the number of flare pixels in one embodiment.
The vertical axis of the graph in FIG. 9 represents the total number of pixels indicating flare detected by the method described using FIG. 6, and the horizontal axis represents time. The graph in FIG. 9 shows, for example, the change in the number of flare pixels calculated based on a flare image taken when the flare became large, and the change in the number of flare pixels calculated based on a flare image taken at a different time when the flare became smaller. The changes in the number of pixels are displayed side by side. A graph 91 shows a change in the number of pixels detected when the flare becomes large, and a graph 92 shows a change in the number of pixels detected when the flare becomes small.
A result was obtained in which the size of the flare visually observed matched the number of pixels of the flare detected by the video analysis unit 33.

Next, a method for estimating the flare gas flow rate will be explained.
FIG. 10 is a diagram illustrating a flare gas flow rate estimation process in one embodiment.
FIG. 10 shows an example of a calibration curve showing the relationship between the number of pixels indicating flare in a flare image and the flow rate of flare gas. The vertical axis of FIG. 10 is the total number of pixels indicating flare detected by the method explained using FIG. 6, and the horizontal axis is the measured value of the gas flow rate measured by the flow meter installed in the flare stack. be. A monitoring system 1 is introduced into a flare stack equipped with a flow meter that measures the flow rate of flare gas, and a flare image is photographed using, for example, a camera 10A. The video analysis unit 33 analyzes the flare video, calculates the number of pixels indicating flare, and records it in the storage unit 35 together with the time when the flare video was photographed. In parallel with this process, information on the flow rate measured by the flow meter is acquired together with the measurement time, and is recorded in a predetermined storage device. When the total number of pixels at each time recorded in the storage unit 35 and the measured value of the flow rate measured at the same time are mapped in correspondence, a calibration curve 101 as illustrated in FIG. 10 is obtained using the least squares method or the like. is obtained. The calibration curve 101 created in this way is recorded in the storage section 35. The flare gas flow rate estimation unit 34 refers to the calibration curve 101 recorded in the storage unit 35 and calculates an estimated value of the flare gas flow rate corresponding to the total number of pixels of the flare detected by the video analysis unit 33. Further, the flare gas flow rate estimation unit 34 compares the estimated value of the flare gas flow rate with a threshold value, and instructs the output unit 36 to issue an alarm if the estimated value of the flare gas flow rate exceeds the threshold value.

(Setting process)
Next, a process for estimating the flare gas flow rate by the monitoring system 1 will be described.
First, prior to estimating the flare gas flow rate, various setting processes are performed.
FIG. 11 is a diagram illustrating an example of setting processing in one embodiment.
First, the image acquisition unit 31 acquires flare images taken by the cameras 10A and 10B (step S1). The output unit 36 displays flare images taken by each camera on the monitor of the monitoring device 30. Next, the observer refers to each flare image and sets the tip position of the burner section 51 (step S2). For example, the observer inputs the position of the tip by inputting coordinate information of the tip. The setting information acquisition unit 32 acquires the input information. When the coordinate information of the tip position is input, the setting information acquisition section 32 records the coordinate information in the storage section 35 as the setting information of the tip position of the burner section 51. The position of the tip of the burner section 51 may be set by the video analysis section 33 through image analysis.

Next, the observer refers to each flare video and sets a flare monitoring area 52 and a background area 53 (step S3). The flare monitoring area 52 is set for the purpose of detecting typical RGB luminance values of pixels in which flare is captured. The observer sets a small area near the tip of the burner section 51 as a flare monitoring area 52. This is to prevent backgrounds other than flare from being included. The observer sets the flare monitoring area 52 by inputting the coordinates of the vertices of the rectangular area corresponding to the flare monitoring area 52, for example. The setting information acquisition unit 32 records the input coordinate information in the storage unit 35 as setting information of the flare monitoring area 52.

The background area 53 is set for the purpose of detecting RGB luminance values of a background such as the sky. The observer sets a background area 53 as a range in which flare and smoke are not reflected. The observer sets the background area 53 by inputting the coordinates of the vertices of the rectangular area corresponding to the background area 53, etc. The setting information acquisition unit 32 records the input coordinate information in the storage unit 35 as setting information of the background area 53.

When these settings are made and the observer performs a predetermined operation, the video analysis unit 33 calculates the RGB luminance value of each pixel included in the flare monitoring area 52 in the flare video acquired by the video acquisition unit 31. . The video analysis unit 33 selects the RGB brightness value of an arbitrary pixel and records this value in the storage unit 35 as a typical flare RGB brightness value (corresponding to point 61 in FIG. 6). Further, the video analysis unit 33 calculates the RGB luminance value of each pixel included in the background area 53 in the flare video acquired by the video acquisition unit 31. The video analysis unit 33 selects the RGB brightness value of an arbitrary pixel and records this value in the storage unit 35 as the background RGB brightness value (corresponding to point 63 in FIG. 6). Alternatively, the RGB brightness value of each pixel included in the flare monitoring area 52 calculated by the video analysis unit 33 is displayed on the display unit of the monitoring device 30, and the watcher can select an appropriate RGB brightness while referring to the displayed brightness value. Representative flare RGB brightness values may be set by inputting the values into the monitoring device 30. The same applies to the RGB luminance values of the background.

Note that a plurality of sets of typical RGB brightness values of the flare and typical RGB brightness values of the background may be set depending on the time of day, season, weather, and the like. For example, the storage unit 35 stores typical RGB luminance values of flare and background RGB luminance values on a sunny day, typical RGB luminance values of flare and background RGB luminance values on a rainy day, and typical RGB luminance values of flare and background RGB luminance values on a rainy day. A typical RGB brightness value and a background RGB brightness value, a typical RGB brightness value of an evening flare and a background RGB brightness value, etc. may be recorded. Furthermore, if there are a plurality of representative RGB brightness values of the flare (for example, an RGB brightness value near the center of the flame and an RGB brightness value near the boundary), a plurality of RGB brightness values may be set.

Next, the monitor sets a threshold value for determining whether to issue an alarm in the monitoring device 30 (step S4). The monitor may set a plurality of threshold values depending on the flow rate of flare gas. For example, if the flow rate of flare gas exceeds "X1" ml/sec, the monitor will issue an alarm saying "Flare gas flow rate has exceeded If it exceeds X2, set it to issue an alarm saying "Flare gas flow rate has exceeded X2." The setting information acquisition unit 32 records the input threshold value and alarm content in the storage unit 35.

(Flare gas flow rate estimation process)
Next, the flare gas flow rate estimation process will be explained. First, a process of estimating the flare gas flow rate using only the flare image captured by one camera 10A will be described.
FIG. 12 is a first diagram illustrating an example of a flare gas flow rate estimation process in one embodiment. First, the video acquisition unit 31 acquires a flare video captured by the camera 10A (step S11). The video acquisition unit 31 outputs the flare video to the video analysis unit 33. The video analysis unit 33 performs the following processing for each pixel of the flare video.
First, the video analysis unit 33 analyzes the RGB luminance values of a pixel in a flare video (step S12). Next, the video analysis unit 33 plots the analyzed RGB luminance values in an RGB three-dimensional space whose coordinate axes are the R luminance value, the G luminance value, and the B luminance value (step S13). The plotted point corresponds to point 62 in FIG. Next, the video analysis unit 33 calculates the distance from the typical flare RGB luminance value and the background RGB luminance value (step S14). Specifically, the video analysis unit 33 calculates the distance between the point corresponding to the RGB luminance value of a typical flare (point 61 in FIG. 6) and the plotted point (point 62 in FIG. 6). The video analysis unit 33 also calculates the distance between the point corresponding to the RGB luminance value of the background (point 63 in FIG. 6) and the plotted point (point 62 in FIG. 6). At this time, for example, if typical RGB brightness values of flare and background RGB brightness values are set for each time period, the video analysis unit 33 determines the RGB brightness values of flare and background according to the current time period. The information is read out from the storage unit 35 and plotted in a three-dimensional space. Further, for example, if a plurality of representative flare RGB brightness values are set, the distance between the point corresponding to each RGB brightness value and the point plotted in step S13 is calculated.

Next, the video analysis unit 33 determines whether the point corresponding to the analyzed RGB brightness value is close to the point corresponding to the typical RGB brightness value of flare (step S15). If the point is close to the point corresponding to the typical RGB luminance value of flare (step S15; Yes), the video analysis unit 33 determines that the currently analyzed pixel is a pixel in which flare is captured (step S16). If the point is close to the point corresponding to the RGB luminance value of the background (step S15; No), the video analysis unit 33 determines that the currently analyzed pixel is a pixel in which the background is reflected (step S17). Next, the video analysis unit 33 determines whether all pixels of the flare video have been analyzed (step S18), and repeats the processing from step S12 until the analysis of all pixels is completed. While repeating the processing of steps S12 to S18, the video analysis unit 33 creates a flare image (FIG. 7(a)) in which pixels determined to be flare and pixels determined to be background are given different colors.

It should be noted that the processes from step S12 to step S18 are performed on all pixels included in the flare image, but areas where flare is clearly not visible, such as areas where flare stack equipment is visible, are excluded from processing. Good too. Alternatively, a monitoring target area may be set, and the processes of steps S12 to S18 may be performed only on pixels within the monitoring target area.

When the process of determining whether all pixels are flare or background is completed, the video analysis unit 33 next determines whether a cavity exists in the group of pixels determined to be flare (step S19). While scanning the flare image in the vertical direction, the video analysis unit 33 analyzes the partial image being scanned in the horizontal direction, and determines the pixel group in which the flare appears, the pixel group in which the background appears, and the pixel group in which the flare appears. When pixel groups are detected in this order, it is determined that a cavity exists. If a cavity exists (step S19; Yes), the cavity complement process described using FIG. 7 is performed. Specifically, the video analysis unit 33 updates the flare image by adding a color indicating that it is a flare to a group of non-flare images sandwiched between a group of flare pixels. If a cavity does not exist (step S19; No), the process proceeds to step S21. Next, the video analysis unit 33 determines whether the cavity detection process has been performed for all partial images (step S21), and repeats the process from step S19 until the cavity detection process and complementation process are completed for all the partial images. repeat.

In addition, flare complementation processing is performed while scanning the flare image horizontally and scanning the partial image being scanned in the vertical direction. If pixel groups are detected in the order of groups, it may be determined that a cavity exists.

When the cavity complementation process is completed, the video analysis unit 33 then calculates the total number of pixels determined to be flare based on the flare image (step S22). The pixels determined to be flare include the pixels determined to be flare in step S16 as well as the pixels complemented in step S20. The video analysis section 33 outputs the calculated number of pixels to the flare gas flow rate estimation section 34. Next, the flare gas flow rate estimating unit 34 calculates the flow rate of the flare gas being emitted from the burner unit 51 based on the number of pixels determined to be flare obtained from the video analysis unit 33 and the calibration curve 101 illustrated in FIG. presume. Further, the flare gas flow rate estimation unit 34 compares the estimated value of the flare gas flow rate with the threshold value set in step S4. When the estimated value of the flare gas flow rate exceeds the threshold value, the flare gas flow rate estimation unit 34 instructs the output unit 36 to output an alarm corresponding to the threshold value. The output unit 36 outputs an estimated value of the flare gas flow rate and an alarm (step S23). Thereby, the flow rate of the flare gas released from the burner section 51 can be estimated based on the flare image taken by the camera 10A.

Next, a process of estimating the flare gas flow rate using two flare images taken by the two cameras 10A and 10B will be described.
FIG. 13 is a second diagram illustrating an example of the flare gas flow rate estimation process in one embodiment. The processing from step S11 to step S21 is the same as the processing described in FIG. 12. First, the image acquisition unit 31 acquires flare images captured by the cameras 10A and 10B (step S11). Next, the video analysis unit 33 performs the processes from step S12 to step S18 for each of the two flare videos. First, the video analysis unit 33 analyzes RGB luminance values for each pixel (step S12), and plots them in an RGB three-dimensional space (step S13). Next, the video analysis unit 33 calculates the distance between the plotted point and the representative RGB luminance value of the flare and the RGB luminance value of the background (step S14), and the RGB luminance value of the pixel currently being analyzed is representative. It is determined whether the RGB luminance value is close to the typical flare RGB luminance value (step S15). If the RGB luminance value is close to the typical flare RGB luminance value (step S15; Yes), the video analysis unit 33 determines that the currently analyzed pixel is a pixel in which flare is captured (step S16). If the RGB luminance value is close to the background RGB luminance value (step S15; No), the video analysis unit 33 determines that the currently analyzed pixel is a pixel in which the background is reflected (step S17).
The video analysis unit 33 performs the above processing on all pixels included in the two flare videos taken by the cameras 10A and 10B, and creates respective flare images.

Next, the video analysis unit 33 determines whether a cavity exists between the pixel groups determined to be flare (step S19), and if there is a cavity, performs a process of interpolating the cavity (step S20).
The video analysis unit 33 performs the above processing on each of the two flare images to complement the cavities. Here, the flare image after the complementation process based on the flare image taken by the camera 10A is called a flare image a, and the flare image after the complementation process based on the flare image taken by the camera 10B is called a flare image b.

Next, the video analysis unit 33 calculates, for the flare image a, the horizontal width W1 and the vertical height H1 of the area in which the flare is captured (step S22A). Here, refer to FIG. 7(b). The horizontal width W1 is the number of pixels in the horizontal direction in the widest part of the flare area. The vertical height H1 is the number of pixels in the vertical direction in the highest part of the flare area.
Further, the video analysis unit 33 calculates the horizontal width W2 and the vertical height H2 of the area in which the flare is captured for the flare image b (step S22A). The method of calculating width W2 and height H2 is the same as W1 and H1.

Next, the video analysis unit 33 calculates the volume V of the rectangular parallelepiped that includes the flare (step S22B). Specifically, when the height H1 and the height H2, whichever has the larger number of pixels, is Max(H1, H2), the video analysis unit 33 calculates the volume V using the following equation (1).
Volume V = Max (H1, H2) x Width W1 x Width W2 (1)
The volume V indicates the number of pixels included in the rectangular parallelepiped containing the flare, and in this embodiment, this is used as an approximate value of the number of pixels of the flare. For example, if the flare image taken from camera 10A appears small (thin) due to the influence of wind, but the flare image taken from another direction shows a wide flare. By calculating the volume V, it is possible to calculate the number of pixels corresponding to the size of the flare that is close to the actual size.

After calculating the volume V, the flare gas flow rate estimating unit 34 estimates the flow rate of the flare gas being emitted from the burner unit 51 based on the volume V (number of pixels) and the calibration curve. Note that a calibration curve created based on the relationship between the volume V and the measured value of the flare gas flow rate is registered in the storage unit 35.
Further, when the estimated value of the flare gas flow rate exceeds the threshold value, the flare gas flow rate estimating unit 34 instructs the output unit 36 to output an alarm corresponding to the threshold value. The output unit 36 outputs an estimated value of the flare gas flow rate and an alarm (step S23).

According to the monitoring system 1 of the present embodiment, the flow rate of flare gas released from the burner section 51 can be estimated simply by inputting flare images to the monitoring device 30. Moreover, by setting a plurality of threshold values, an alarm can be issued in real time when the flare gas flow rate exceeds the threshold value. This allows the amount of flare gas loss to be ascertained without the need for uncertain visual confirmation of flares or the introduction of expensive flow meters. Further, as shown in FIG. 1, while performing conventional monitoring by displaying a flare image on the monitoring monitor 25, it is possible to refer to the estimation result of the flare gas flow rate notified by the monitoring device 30 and the reported alarm. In addition, the observer can arbitrarily set the flare monitoring area 52, background area 53, threshold value, brightness value corresponding to flare, etc., depending on the nature of the surplus gas and the installation environment of the flare stack. It can also be applied to such equipment.

Each process in the monitoring device 30 can be realized by, for example, a CPU included in the monitoring device 30 executing a program. 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 the recording medium. Note that the monitoring device 30 includes hardware such as an OS and peripheral devices. Further, the computer-readable recording medium is, for example, a portable medium such as a flexible disk, magneto-optical disk, ROM, or CD-ROM, or a storage device such as a hard disk built into the monitoring device 30. In addition, computer-readable recording media may dynamically retain programs for short periods of time, such as communication lines used to transmit programs over networks such as the Internet or communication lines such as telephone lines. It may also include a device that retains 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. Further, the above-mentioned program may be one for realizing a part of the above-mentioned functions, or may be one that can realize the above-mentioned functions 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 spirit of the present invention. Further, the technical scope of the present invention is not limited to the above-described embodiments, and various changes can be made without departing from the spirit of the present invention.
Note that RGB luminance values are values expressed by a combination of red (R), green (G), and blue (B) luminance values (0 to 255) in the RGB color space, and black and white luminance values. is a value expressed in 256 gradations from 0 (black) to 255 (white).

1... Surveillance system 10A, 10B... Camera 15... Video acquisition device 20... Video distribution device 25... Monitor for monitoring 30... Monitoring device 31... Video acquisition unit 32... -Setting information acquisition section 33...Video analysis section 34...Flare gas flow rate estimation section 35...Storage section 36...Output section 50...Flare video 52...Flare monitoring area 53...Background area

Claims (7)

an image acquisition unit that acquires an image of the flare;
Based on the luminance value of a predetermined flare pixel in which flare is captured and the luminance value of a predetermined background pixel in which non-flare is captured, each pixel included in the video is divided into a first pixel close to the flare pixel and the background. a video analysis unit that classifies the pixel as any second pixel close to the pixel and calculates the total number of pixels classified as the first pixel;
a flare gas flow rate estimation unit that estimates the flow rate of flare gas , which is gas flowing through the flare gas piping, based on the total number of pixels;
an output unit that compares the flow rate of the flare gas with a predetermined threshold value and issues an alarm if the flow rate of the flare gas exceeds the threshold value ;
Equipped with
The image acquisition unit acquires the images taken from two different directions,
The video analysis unit calculates the number W1 of pixels in the horizontal direction and the number W1 of pixels in the vertical direction of the pixel group classified as the first pixel based on the video shot from a first direction among the acquired video. A number H1 is calculated, and a number W2 of pixels in the horizontal direction and a number H2 of pixels in the vertical direction of the pixel group classified as the first pixel are calculated based on the image taken from the second direction, Calculating the volume V by multiplying the larger number of pixels of the number of pixels H1 and the number of pixels H2, the number of pixels W1, and the number of pixels W2,
The flare gas flow rate estimation unit estimates the flow rate of the flare gas based on the total number of pixels, with the volume V being the total number of pixels.
Monitoring equipment. When one pixel included in the video is taken as a detection target pixel, the video analysis unit analyzes the RGB brightness values of the detection target pixel, and converts the analyzed RGB brightness values into coordinate axes of each RGB brightness value. first coordinate information plotted in a three-dimensional space, second coordinate information plotting an RGB luminance value of the flare pixel in the three-dimensional space, and RGB luminance value of the background pixel plotted in the three-dimensional space. a first distance indicating a distance between the first coordinate information and the second coordinate information using the plotted third coordinate information; and a first distance indicating the distance between the first coordinate information and the second coordinate information, and the first coordinate information and the third coordinate information. and a second distance indicating the distance between the pixel and the target pixel, and whether the detection target pixel is classified as the first pixel or the second pixel based on the first distance and the second distance. determine whether
The monitoring device according to claim 1. The video analysis unit sets an RGB brightness value of the flare pixel and an RGB brightness value of the background pixel according to the time of day or the weather.
The monitoring device according to claim 2. When the detection target pixel classified as the second pixel exists in a position sandwiched between the pixel group classified as the first pixel, the video analysis unit converts the detection target pixel into the first pixel. classified into,
The monitoring device according to claim 2 or 3. The flare gas flow rate estimation unit estimates the flare gas flow rate based on a calibration curve that defines a relationship between the flare gas flow rate and the total number of pixels.
A monitoring device according to any one of claims 1 to 4. two imaging devices that take images of the area around the flare emission part of the flare stack from different directions;
The monitoring device according to any one of claims 1 to 5 , which acquires the video;
A monitoring system equipped with a step of obtaining a video of the flare;
Based on the luminance value of a predetermined flare pixel in which flare is captured and the luminance value of a predetermined background pixel in which non-flare is captured, each pixel included in the video is divided into a first pixel close to the flare pixel and the background. a step of classifying the pixel into any second pixel close to the pixel and calculating the total number of pixels classified as the first pixel;
estimating the flow rate of flare gas, which is gas flowing through the flare gas piping , based on the total number of pixels;
Comparing the flow rate of the flare gas with a predetermined threshold value, and issuing an alarm if the flow rate of the flare gas exceeds the threshold value ;
has
In the acquiring step, acquiring the images taken from two different directions,
In the step of calculating, the number of pixels in the horizontal direction W1 and the number of pixels in the vertical direction of the pixel group classified as the first pixel are determined based on the image taken from the first direction among the acquired images. A number H1 is calculated, and a number W2 of pixels in the horizontal direction and a number H2 of pixels in the vertical direction of the pixel group classified as the first pixel are calculated based on the image taken from the second direction, Calculating the volume V by multiplying the larger number of pixels of the number of pixels H1 and the number of pixels H2, the number of pixels W1, and the number of pixels W2,
In the estimating step, the volume V is the total number of pixels, and the flow rate of the flare gas is estimated based on the total number of pixels.
Monitoring method.

Images