Gas leakage monitoring device, gas leakage monitoring method, and program

Abstract

The present invention provides a gas leak monitoring device comprising: an image data acquisition unit that acquires image data generated by taking pictures at different times as raw image data; a vector generation unit that selects two raw image data in chronological order from the raw image data and performs predetermined image processing on the selected two raw image data to calculate the movement direction and movement speed of objects contained in the images, thereby generating a vector representing the movement direction and movement speed for each pixel contained in the raw image data; and a frequency distribution generation unit that classifies the vectors generated by the vector generation unit according to a predetermined vector length range to generate a frequency distribution for gas leak monitoring.

Description

Technical Field

[0001] This invention relates to a gas leak monitoring device, a gas leak monitoring method, and a procedure. This application claims priority based on Japanese Patent Application No. 2023-204587, filed on December 4, 2023, the contents of which are incorporated herein by reference. Background Technology

[0002] Gases such as CO2 have unique absorption wavelength bands, and a gas leak detection method utilizing the characteristics of these absorption wavelength bands is known (see, for example, Patent Document 1). For instance, CO2 gas has an absorption wavelength band around 4.3 μm. Therefore, when photographed by an infrared camera equipped with a filter that allows infrared light to pass through at 4.3 μm, the following phenomenon occurs: If CO2 gas is present in a region between the object being photographed and the infrared camera, an intensity difference arises between the electromagnetic waves radiated from the object and directly reaching the infrared camera and the electromagnetic waves that reach the infrared camera through the CO2 gas, due to absorption caused by the CO2 gas. This intensity difference of electromagnetic waves manifests as a difference in brightness values ​​in the image data generated by the infrared camera, thereby enabling the visualization of the CO2 gas.

[0003] As a method for detecting CO2 gas visualized in image data, for example, Patent Document 1 discloses the following method: In image data generated by an infrared camera, there exists a difference in brightness values ​​as described above. Pixels whose brightness value change per unit time is below a predetermined threshold, using this difference in brightness values, are identified as pixels containing gas.

[0004] Previous technical documents

[0005] Patent documents

[0006] Patent Document 1: Japanese Patent No. 6665863 Summary of the Invention

[0007] The technical problem to be solved by the invention

[0008] However, there is a desire to determine the presence or absence of gas leaks with high accuracy and in the shortest possible time. From this perspective, the method disclosed in Patent Document 1, which involves processing the change in brightness value for each pixel within a unit of time, presents the following problem: To improve the accuracy of the change in brightness value per unit of time, the number of images captured needs to be increased; however, increasing the number of images leads to a longer processing time.

[0009] The present invention was made to solve the above-mentioned problems, and its purpose is to provide a gas leak monitoring device, gas leak monitoring method and procedure that can indicate whether there is a gas leak with high accuracy and in a shorter time.

[0010] means for solving technical problems

[0011] To address the aforementioned issues, the gas leak monitoring device of the present invention comprises: an image data acquisition unit that acquires image data generated by taking pictures at different times as raw image data; a vector generation unit that selects two raw image data in chronological order from the raw image data and performs predetermined image processing on the selected two raw image data to calculate the movement direction and speed of objects contained in the images, thereby generating a vector representing the movement direction and speed for each pixel contained in the raw image data; and a frequency distribution generation unit that classifies the vectors generated by the vector generation unit according to a predetermined vector length range to generate a frequency distribution for gas leak monitoring.

[0012] The gas leak monitoring method of the present invention includes the following steps: acquiring image data generated by taking pictures at different times as raw image data; selecting two raw image data in chronological order from the acquired raw image data, and performing predetermined image processing on the selected two raw image data to calculate the movement direction and speed of the objects contained in the images, thereby generating a vector representing the movement direction and speed for each pixel contained in the raw image data; and classifying the generated vectors according to a predetermined vector length range to generate a frequency distribution for gas leak monitoring.

[0013] The program involved in this invention enables a computer to perform the following steps: acquiring image data generated by taking pictures at different times as raw image data; selecting two raw image data in chronological order from the acquired raw image data, and performing predetermined image processing on the selected two raw image data to calculate the direction and speed of movement of objects contained in the images, thereby generating a vector representing the direction and speed of movement for each pixel contained in the raw image data; and classifying the generated vectors according to a predetermined vector length range to generate a frequency distribution for gas leak monitoring.

[0014] Invention Effects

[0015] The gas leak monitoring device, gas leak monitoring method and procedure according to the present invention can indicate whether there is a gas leak with high accuracy and in a shorter time. Attached Figure Description

[0016] Figure 1 This is a block diagram illustrating a structural example of a gas leak monitoring system according to an embodiment of the present invention.

[0017] Figure 2 This is a flowchart illustrating an example of the operation of the image data acquisition unit of the gas leak monitoring device according to an embodiment of the present invention.

[0018] Figure 3 This is a flowchart illustrating an example of the operation of the image processing unit, vector generation unit, filtering unit, frequency distribution generation unit, and determination unit included in the gas leak monitoring device according to an embodiment of the present invention.

[0019] Figure 4 This is a flowchart of a subroutine for image processing performed by the image processing unit of the gas leak monitoring device according to an embodiment of the present invention.

[0020] Figure 5 This is a diagram illustrating an example of an image represented by raw image data according to an embodiment of the present invention.

[0021] Figure 6 This is a diagram illustrating an example of an image represented by processed image data according to an embodiment of the present invention.

[0022] Figure 7 This is a diagram illustrating an example of the original image data and the first vector generated based on the original image data in an embodiment of the present invention.

[0023] Figure 8 This is a diagram illustrating an example of processed image data and a second vector generated based on the processed image data, as described in an embodiment of the present invention.

[0024] Figure 9 This is a diagram showing a general outline of filtering performed by the filtering unit according to an embodiment of the present invention.

[0025] Figure 10 This diagram illustrates an example of processed image data and vectors extracted by filtering by the filtering unit according to an embodiment of the present invention.

[0026] Figure 11 This is an example showing a histogram corresponding to the frequency distribution generated by the frequency distribution generation unit according to an embodiment of the present invention.

[0027] Figure 12 It is a schematic block diagram showing the structure of a computer according to at least one embodiment. Detailed Implementation

[0028] Hereinafter, with reference to the figures, the gas leak monitoring device, gas leak monitoring method, and procedure according to embodiments of the present invention will be described. Furthermore, in the figures, the same or corresponding structures are represented by the same symbols, and descriptions are omitted where appropriate.

[0029] (System Structure)

[0030] Figure 1 This is a block diagram illustrating the structure of a gas leak monitoring system 1 according to an embodiment of the present invention. The gas leak monitoring system 1 includes a camera device 2, a communication network 3, and a gas leak monitoring device 10. In the gas leak monitoring system 1, the gas to be monitored is CO2 gas. The communication network 3 is, for example, a communication network operated by a telecommunications operator, and is a communication network capable of both wired and wireless connections.

[0031] The camera device 2 is, for example, an infrared camera equipped with a filter that transmits infrared light at a wavelength of 4.3 μm, which is the absorption wavelength band of CO2 gas. The object photographed by the camera device 2 is, for example, a site equipped with a complete set of equipment. The camera device 2 is, for example, mounted on an unmanned aerial vehicle (UAV), and while moving with the UAV, it captures dynamic images of the site from above. The camera device 2 is equipped with a communication device that wirelessly connects to the communication network 3, and transmits the dynamic image data generated by the capture to the gas leak monitoring device 10 via the communication network 3.

[0032] (Structure of a gas leak monitoring device)

[0033] The gas leak monitoring device 10 can be constructed using computers such as servers, personal computers, and microcomputers, as well as peripheral devices of such computers. As a functional structure composed of hardware such as the computer and software such as programs executed by the computer, it includes an image data acquisition unit 11, a raw image data storage unit 12, an image processing unit 13, a processed image data storage unit 14, a vector generation unit 15, a filtering unit 16, a frequency distribution generation unit 17, a determination unit 18, and a display unit 19.

[0034] The image data acquisition unit 11 is connected to the communication network 3 via a wired or wireless connection and receives moving image data transmitted by the camera device 2. The moving image data received by the image data acquisition unit 11 contains frames arranged in the order of capture, and time information indicating the time of capture is associated with each frame. The image data acquisition unit 11 divides the frames contained in the received moving image data into a time sequence. The image data acquisition unit 11 generates original image data from each of the divided frames, associates the generated original image data with the time information associated with the corresponding frame, and records it in the original image data storage unit 12.

[0035] The image processing unit 13 performs predetermined image processing on the original image data recorded in the original image data storage unit 12 to generate processed image data corresponding to the original image data. The image processing unit 13 associates the generated processed image data with the time information associated with the original image data corresponding to the processed image data and records it in the processed image data storage unit 14.

[0036] The vector generation unit 15 selects two raw image data that are adjacent in time sequence from the raw image data storage unit 12. The vector generation unit 15 then selects two processed image data corresponding to each of the two selected raw image data. Specifically, the two processed image data corresponding to each of the two raw image data are two processed image data that are associated with the same time information as the two raw image data, and the two processed image data are also adjacent in time sequence.

[0037] The vector generation unit 15 performs predetermined image processing on the two selected original image data, calculating the direction and speed of movement of the objects contained in the images, and generates a vector representing the direction and speed of movement for each pixel (hereinafter referred to as the first vector). The vector generation unit 15 performs predetermined image processing on the two selected processed image data, calculating the direction and speed of movement of the objects contained in the images, and generates a vector representing the direction and speed of movement for each pixel (hereinafter referred to as the second vector).

[0038] Image processing, such as optical flow processing in the Gunnar Farneback method, is used to calculate the direction and speed of movement of objects contained in an image.

[0039] The filtering unit 16 performs a filtering process on the second vector of each pixel, using a first vector that corresponds to each pixel in the second vector. This process emphasizes and extracts the vector corresponding to the gas. The frequency distribution generation unit 17 generates a frequency distribution by classifying the vectors of each pixel output by the filtering unit 16 according to a preset vector length range. The determination unit 18 determines whether there is a gas leak based on the frequency of the preset interval in the frequency distribution generated by the frequency distribution generation unit 17 and a preset threshold.

[0040] The display unit 19 is, for example, a display device such as a liquid crystal display, which displays the frequency distribution generated by the frequency distribution generation unit 17 in the form of a histogram, or displays the determination result of the determination unit 18.

[0041] (Example of the operation of a gas leak monitoring device)

[0042] refer to Figures 2 to 4 The flowchart illustrates the operation of the gas leak monitoring device 10.

[0043] like Figure 2 As shown, the image data acquisition unit 11 receives motion image data (Sa1) sent by the imaging device 2. The image data acquisition unit 11 continues to receive motion image data sent by the imaging device 2, and sequentially segments the data starting from the most recently received frame, generating image data based on the segmented frames. The image data acquisition unit 11 performs noise removal processing on the generated image data. Here, the motion image data captured by the imaging device 2 is color motion image data. The image data acquisition unit 11 converts the image data into grayscale image data to preserve the brightness of the noise-removed image data, thereby generating original image data. The pixel value of each pixel in the original image data becomes a value representing the brightness value, expressed as any value from 0 to 255 (Sa2).

[0044] Figure 5 The image shown is an example of an image represented by the original image data. Figure 5 In this context, the object represented by the symbol 100 is, for example, a structure that is fixedly installed in a complete set of equipment.

[0045] The image data acquisition unit 11 associates the raw image data with the time information associated with the frame from which the raw image data was generated and records this association in the raw image data storage unit 12 (Sa3). The image data acquisition unit 11 determines whether to continue receiving moving image data (Sa4). If it is determined that moving image data should continue to be received (Sa4 is "Yes"), the image data acquisition unit 11 performs Sa2 processing on frames that are not segmented. On the other hand, if it is determined that moving image data should not continue to be received (Sa4 is "No"), the image data acquisition unit 11 ends the processing.

[0046] Furthermore, the image data acquisition unit 11 can perform Sa4 determination processing after segmenting one frame and processing Sa2 and Sa3, or it can segment a certain number of frames, process each segmented frame for Sa2 and Sa3, and then perform Sa4 determination processing. Thus, the raw image data storage unit 12 stores raw image data arranged in chronological order and associated with time information.

[0047] like Figure 3 As shown, the image processing unit 13 determines whether a predetermined amount of raw image data (Sb1) is recorded in the raw image data storage unit 12. Here, the predetermined amount is the amount of raw image data required for the predetermined image processing of the emphasized gas performed by the image processing unit 13, for example, a value of about "5" or "6" is preset.

[0048] If it is determined that a predetermined number of original image data are not recorded in the original image data storage unit 12 (Sb1 is "No"), the image processing unit 13 performs the processing of Sb1 again. On the other hand, if it is determined that a predetermined number of original image data are recorded in the original image data storage unit 12 (Sb1 is "Yes"), the image processing unit 13 reads the predetermined number of recorded original image data and the time information associated with the original image data from the original image data storage unit 12. The image processing unit 13 uses the read original image data as reference original image data (Sb2-1). If the reference original image data is determined, the image processing unit 13 begins the aforementioned predetermined image processing emphasizing the gas, i.e. Figure 4 The image processing subroutine (Sb3) shown.

[0049] (Pre-defined image processing of the gas, performed by the image processing department)

[0050] The image processing unit 13 records the time information corresponding to the reference original image data, which is read along with the reference original image data, in its internal storage area (Sc1). The image processing unit 13 reads (Sc2) a number of original image data that are earlier than the time shown in the time information corresponding to the reference original image data and are consecutive in time sequence. The processing of Sc2 is performed after the processing of Sb1, so when the processing of Sc2 is performed, at least (Sc2) a number of original image data that are earlier than the reference original image data in time sequence are stored in the original image data storage unit 12.

[0051] Here, the definitions of the terms "earlier time" and "later time" are explained. For example, suppose there are three original image data sets associated with three time information sets: "0:00:01", "0:00:02", and "0:00:03". In this case, the original image data associated with the time information set at a time earlier than the time information set with the original image data set at "0:00:02" is called the original image data set at "0:00:01". Conversely, the original image data associated with the time information set with the time information set with the original image data set at a time later than the time information set with the original image data set at "0:00:02" is called the original image data set at "0:00:03". The definitions of the terms "earlier time" and "later time" are the same below.

[0052] The image processing unit 13 generates a combination (Sc3) of two original image data arranged sequentially in time series order based on a predetermined number of original image data obtained by adding reference original image data to the read original image data. Here, sequentially arranged in time series order means that the time series order does not have to be consecutive, as long as they are arranged sequentially in time series order. Therefore, the image processing unit 13 generates a predetermined number C2 combinations.

[0053] The image processing unit 13 generates differential image data for each combination by performing the following processing for each combination. The image processing unit 13 performs the following processing: reads the brightness values ​​of pixels at the same position in the two original image data contained in a certain combination, and uses the absolute value of the difference between the two read brightness values ​​as the brightness value of the pixel at that position in the differential image data. The image processing unit 13 performs this processing on all pixels to generate differential image data (Sc4) corresponding to that combination.

[0054] The image processing unit 13 performs binarization of the brightness values ​​of each pixel in each generated combination of differential image data. The image processing unit 13 performs the following binarization process (Sc5): the brightness value of each pixel in the differential image data is compared with a preset threshold. For example, if the brightness value is above the threshold, the brightness value of the pixel is set to "1", and if the brightness value is below the threshold, the brightness value of the pixel is set to "0".

[0055] The image processing unit 13 accumulates the brightness value for each pixel of all the binarized differential image data as shown in the following formula (1), and calculates the brightness change frequency (Sc6) as the sum of the accumulated brightness values.

[0056] [Formula 1]

[0057]

[0058] In equation (1), S(x, y) is the cumulative brightness change frequency of pixel (x, y), and lt(x, y) is the brightness value of pixel (x, y) in the t-th differential image data. The image processing unit 13 detects the maximum value of the brightness change frequency S(x, y) in the brightness change frequency distribution representing the brightness change frequencies of all generated pixels, and normalizes the brightness change frequency S(x, y) of each pixel in the brightness change frequency distribution so that the maximum value becomes a predetermined value (e.g., "255"). Specifically, the image processing unit 13 normalizes the brightness change frequency S(x, y) of each pixel in the brightness change frequency distribution by dividing it by the maximum value of the brightness change frequency S(x, y) of the brightness change frequency distribution and multiplying it by the predetermined value (Sc7).

[0059] For example, with a predetermined value of "255", the pixel values ​​included in the normalized brightness variation frequency distribution become values ​​between 0 and 255. If these pixel values ​​are considered as brightness values, the normalized brightness variation frequency distribution becomes grayscale image data containing the same number of pixels as the original image data. Therefore, the normalized brightness variation frequency distribution will be referred to as the processed image data below.

[0060] The image processing unit 13 reads out the time information associated with the reference raw image data stored in the internal storage area. The image processing unit 13 associates the read-out time information with the processed image data and records it in the processed image data storage unit 14 (Sc8), and then ends the subroutine processing.

[0061] Figure 6 The image shown is from the Figure 5 An example of an image represented by processed image data obtained using the original image data as the baseline original image data. Figure 6 In the diagram, the pixels whose brightness values ​​are represented by the symbol 101 are pixels representing gas. It can be seen that in... Figure 5 In the original image data, the gas, which is not clear due to the small difference in brightness between itself and the background, is... Figure 6 The processed image data shown is highlighted and becomes clearer.

[0062] Return to Figure 3The image processing unit 13 determines whether two or more processed image data are stored in the processed image data storage unit 14 (Sb4). If it is determined that no two or more processed image data are stored in the processed image data storage unit 14 (Sb4 is "No"), the image processing unit 13 reads the original image data associated with the time information indicating a time one time after the time shown in the time information stored in the internal storage area, and the time information associated with the original image data from the original image data storage unit 12. The image processing unit 13 uses the read original image data as the reference original image data (Sb2-2). If the reference original image data is determined, the image processing unit 13 starts processing as Sb3 again. Figure 4 The image processing subroutine shown.

[0063] If it is determined that there are two or more processed image data stored in the processed image data storage unit 14 ("Yes" in Sb4), the image processing unit 13 outputs the time information associated with the most recently recorded processed image data in the processed image data storage unit 14 to the vector generation unit 15 (Sb5).

[0064] The vector generation unit 15 reads the time information output by the image processing unit 13. The vector generation unit 15 reads the original image data corresponding to the read time information and the original image data that is adjacent in time sequence to the original image data and corresponds to a time earlier than the time indicated by the read time information from the original image data storage unit 12. The vector generation unit 15 performs optical flow processing on the two read original image data to generate a vector representing the movement direction and speed of each pixel, namely the first vector (Sb6). Furthermore, the processing of Sb6 is performed after the processing of Sb1 and Sb4; therefore, when processing Sb6, at least one original image data at a time earlier than the original image data corresponding to the time information is stored in the original image data storage unit 12.

[0065] Figure 7 The image shown is for making the image accessible to the user. Figure 5 The first vector obtained by optical flow processing of the original image data shown and the original image data adjacent to it in time sequence overlaps with the original image data. Figure 5 The image shown is derived from the original image data. Additionally, the first vector is obtained for each pixel, but... Figure 7 In order to make the first vector easier to observe, the first vector is shown with each pixel spaced at a certain interval.

[0066] The vector generation unit 15 reads the processed image data corresponding to the read-in time information and the processed image data that is adjacent to the processed image data in time sequence and corresponds to a time earlier than the time shown in the read-in time information from the processed image data storage unit 14. In this way, by reading the processed image data, the vector generation unit 15 reads out two processed image data that are associated with the same time information as the two original image data read out by itself in the processing of Sb6, that is, two processed image data corresponding to each of the two original image data from the processed image data storage unit 14.

[0067] The vector generation unit 15 performs the same optical flow processing as the processing performed in Sb6 on the two read-out processed image data to generate a vector, namely the second vector, representing the moving direction and moving speed of each pixel.

[0068] The vector generation unit 15 associates the generated second vector with the first vector generated in the processing of Sb6 with information indicating the position of the corresponding pixel, and outputs the information to the filtering unit 16 (Sb7). In addition, the processing of Sb7 is performed after the processing of Sb4. Therefore, when processing of Sb7 is performed, at least one piece of processed image data from a time earlier than the processed image data corresponding to the time information is stored in the processed image data storage unit 14.

[0069] Figure 8 The image shown is for making the image accessible to the user. Figure 6 The processed image data shown, and the second vector obtained by processing the processed image data adjacent to the processed image data in time sequence through optical flow processing, overlap. Figure 6 The image shown is an image derived from the processed image data. Additionally, the second vector is obtained for each pixel, but... Figure 8 To facilitate observation of the second vector, it is shown that each pixel is spaced at regular intervals. From Figure 7 The image shown and Figure 8 As can be seen from the image shown, in Figure 7 In the diagram, the first vector representing the direction and velocity of gas movement is almost never shown, but... Figure 8 The diagram shows a second vector that represents the direction and speed of gas movement.

[0070] The filtering unit 16 reads the associated first and second vectors established from the pixel position information output by the vector generation unit 15. For each pixel, the filtering unit 16 performs the following processing: selects the first and second vectors that are consistent with the pixel, and filters the second vector using the first vector. For example, ... Figure 9As shown, the first vector in pixel (x, y) is the vector Vorg(x, y) represented by the solid arrowhead (symbol 201), and the second vector in the same pixel (x, y) is the vector Vprocessed(x, y) represented by the dashed arrowhead (symbol 202). Here, "θ" represents the angle between vector Vorg(x, y)201 and vector Vprocessed(x, y)202. θ is an angle between 0° and 180°.

[0071] At this time, the filtering unit 16 performs the following filtering: it calculates a vector Vhistogram(x,y) with a length calculated by the following formula (2) and a vector Vhistogram(x,y)204 with the same direction as the vector Vprocessed(x,y)202, which is the second vector. Wherein, when the length calculated by the formula (2) is negative, the direction of the vector Vhistogram(x,y)204 is opposite to the direction of the vector Vprocessed(x,y)202.

[0072] [Formula 2]

[0073]

[0074] If we use vector pairs to explain the filtering performed by the filtering unit 16, it is as follows. The filtering unit 16 calculates the multiplication value obtained by multiplying the norm of the vector Vorg(x,y)201, which is the first vector, as shown in the second term on the right side of equation (2), by cos(θ / 2). Figure 9 As shown, the filtering unit 16 sets the calculated multiplication value as the length and determines a vector 203 that is in the opposite direction to the vector Vprocessed(x,y)202, which is the second vector. The filtering unit 16 adds the determined vector 203 to the vector Vprocessed(x,y)202 to calculate the vector Vhistogram(x,y)204.

[0075] Figure 7 The first vector shown is not clearly represented by gas in the original image data. Therefore, it mainly represents the direction and speed of movement of the calculation object point corresponding to a fixed object such as a structure, which is the object of optical flow calculation. Furthermore, the fixed object is not limited to structures such as those installed in equipment sets, but also includes objects such as soil, sand, and grass. The vector of the calculation object point corresponding to the fixed object is generated by capturing images while the camera device 2 is moving. Therefore, the first vector largely depends on the direction and speed of movement of the camera device 2.

[0076] In contrast, Figure 8The second vector shown is a vector that includes a vector representing the direction and speed of movement of the calculation point corresponding to the fixed object and a vector representing the direction and speed of movement of the calculation point corresponding to the diffusing gas. The direction and speed of gas diffusion are also affected by the pressure applied to the gas, the direction of gas ejection, the strength and direction of the wind around the gas, etc. Therefore, the vector representing the direction and speed of gas movement in the second vector is a vector composed of a vector that depends on the direction and speed of movement of the camera device 2 and a vector generated by influences other than the direction and speed of movement of the camera device 2.

[0077] The purpose of filtering performed by the filtering unit 16 is to remove the vector of the computational object point corresponding to the fixed object, and to emphasize the vector of the computational object point corresponding to the gas. In order to achieve the former purpose of removing the vector of the computational object point corresponding to the fixed object, it is only necessary to subtract the first vector from the second vector of the same pixel. However, in the first vector, if there is a computational object point corresponding to the fixed object in the pixel where gas exists, simply subtracting the first vector from the second vector will result in a decrease in the length of the vector of the computational object point corresponding to the gas. Therefore, in equation (2), the norm of the vector Vorg(x,y) which is the first vector is multiplied by cos(θ / 2).

[0078] With θ = 0°, the first and second vectors point in the same direction, therefore it is inferred that these two vectors are the vectors of the computational point corresponding to the fixed object. With θ = 0°, cos(θ / 2) becomes "1", so the first vector can be directly subtracted from the second vector. Therefore, it is possible to remove the vector of the computational point corresponding to the fixed object from the second vector.

[0079] The hypothesis is as follows: As the angle θ between the first and second vectors increases, the element of the vector corresponding to the gas-related computational object point in the second vector increases. cos(θ / 2) decreases as θ increases, becoming "0" at θ = 180°. Therefore, as the element of the vector corresponding to the gas-related computational object point in the second vector increases, the influence of the first vector can be reduced, and the vector of the gas-related computational object point contained in the second vector can be emphasized. Therefore, the vector Vhistogram(x, y) 204 calculated by the filtering unit 16 for each pixel can be said to be a vector that removes the vector of the computational object point corresponding to the fixed object from the second vector while emphasizing and extracting the vector of the computational object point corresponding to the gas.

[0080] In addition, in the following description, the vector of the calculation object point corresponding to the fixed object will be simply referred to as the vector corresponding to the fixed object, and the vector of the calculation object point corresponding to the gas will be simply referred to as the vector corresponding to the gas.

[0081] Figure 10 The image shown is for passing through the filter unit 16. Figure 7 The first vector shown Figure 8 The second vector shown is calculated using the filter of equation (2), and the vector Vhistogram(x,y)204 for each pixel overlaps with the vector Vhistogram(x,y)204. Figure 6 The image shown is an image derived from processed image data. From... Figure 10 As can be seen from the image shown, by filtering by the filter unit 16, the vectors near the area where the structure exists are removed, and the vectors near the area where the main extracted gas exists are removed.

[0082] The filtering unit 16 associates the vector of each pixel calculated through filtering with information representing the pixel's position. Regarding optical flow processing, for pixels at the outer edges of the image, it's possible to calculate vectors with large detection errors and low reliability in terms of movement speed. Furthermore, in... Figure 10 In the image shown, as in the region indicated by symbol 110, the outer edges of the image sometimes contain information such as an index representing the image. Therefore, the filtering unit 16 excludes vectors from the outer edges based on information indicating the position of pixels associated with the vector. The filtering unit 16 outputs the vector excluding the vectors from the outer edges to the frequency distribution generation unit 17 (Sb8).

[0083] The frequency distribution generation unit 17 reads the vectors output by the filtering unit 16. The frequency distribution generation unit 17 categorizes the read-in vectors according to preset vector length intervals and counts the number of vectors in each interval. The frequency distribution generation unit 17 calculates the frequency of each interval using the ratio of the number of vectors in each interval as the numerator and the total number of read-in vectors as the denominator, and generates a frequency distribution based on the frequency of each interval. The frequency distribution generation unit 17 generates a histogram based on the generated frequency distribution data and displays the generated histogram on the display unit 19.

[0084] Figure 11 This is an example of a histogram displayed on the display unit 19 by the frequency distribution generation unit 17. Figure 11 In the diagram, the vertical axis represents the frequency, and the horizontal axis represents the interval of the vector length range preset in the frequency distribution generation unit 17. Figure 11 In the example shown, the length range of the vector preset in the frequency distribution generation unit 17 is an example of an interval with a length of "0.2".

[0085] The frequency distribution generation unit 17 outputs the generated frequency distribution data to the determination unit 18 (Sb9). The determination unit 18 reads the frequency distribution data output by the frequency distribution generation unit 17. In the determination unit 18, a preset range is established as an indicator for determining whether a gas leak has occurred, and... Figure 11 (a) and Figure 11 (b) The threshold is represented by the symbol 300. Here, the range that serves as the indicator for determining whether there is a gas leak is the range above the length of a vector that is predetermined based on empirical values, etc. In this case, it is set to the range above the length of the vector, which is "999.8".

[0086] The determination unit 18 detects the frequency of each interval in the frequency distribution data whose vector length is "999.8" or higher, and determines whether any of the detected frequencies is above a threshold. If the determination unit 18 determines that any of the detected frequencies is above the threshold, it displays the string "Gas leak present" overlaid on the histogram displayed on the display unit 19 as a determination result, for example. If the determination unit 18 determines that none of the detected frequencies are above the threshold, it displays the string "No gas leak present" overlaid on the histogram displayed on the display unit 19 as a determination result, for example.

[0087] Figure 11 (a) is an example of a histogram taken when images of a location where no gas leak has occurred. Figure 11 (b) is an example of a histogram taken when images of the location of a gas leak were taken. Figure 11 In (a), it is assumed that the frequency of each interval with a vector length of "999.8" or higher does not exceed the threshold represented by the symbol 300. In contrast, in Figure 11 In (b), at least the frequency of the interval “999.8–1000.0” exceeds the threshold represented by the symbol 300. Therefore, in Figure 11 In case (a), the determination unit 18 determines that none of the detected frequencies are above the threshold and displays "No gas leak" on the display unit 19. Figure 11 In case (b), the determination unit 18 determines that any of the detected frequencies is above the threshold and displays "Gas leak" on the display unit 19.

[0088] After making a determination, the determination unit 18 outputs an instruction signal indicating that processing should continue to the image processing unit 13 (Sb10). If the image processing unit 13 receives the instruction signal indicating that processing should continue from the determination unit 18, it determines whether new original image data associated with time information that is later than the time information of the reference original image data stored in the internal storage area should be stored in the original image data storage unit 12 (Sb11).

[0089] If it is determined that new original image data is stored in the original image data storage unit 12 (Sb11 is "Yes"), the image processing unit 13 performs the processing of Sb2-2 again. If it is determined that no new original image data is stored in the original image data storage unit 12 (Sb11 is "No"), the image processing unit 13 ends the processing.

[0090] (Effects)

[0091] In the gas leak monitoring device 10 of the above embodiment, the image data acquisition unit 11 divides the frames of the dynamic image captured by the camera device 2 while it is moving to generate raw image data. The image processing unit 13 performs predetermined image processing on the raw image data to emphasize the gas and generate processed image data. The vector generation unit 15 selects two raw image data that are adjacent in time sequence and two processed image data corresponding to the two raw image data, and performs predetermined image processing on the combination of raw image data and the combination of processed image data to calculate the moving direction and moving speed of the objects contained in the image, generating a first vector and a second vector for each pixel. The filtering unit 16 uses the corresponding first vector in each pixel to emphasize the corresponding second vector and extract the vector corresponding to the gas. The frequency distribution generation unit 17 classifies the extracted vectors into intervals to generate a frequency distribution. The determination unit 18 uses the frequency of each of the preset intervals (e.g., the interval above the length of the preset vector) and a preset threshold to determine whether there is a gas leak in each image.

[0092] Because the gas is not clearly represented in the original image data, the first vector generated from the original image data mainly corresponds to fixed objects such as structures. In contrast, in the processed image data, the gas is emphasized, so the second vector generated from the processed image data includes both vectors corresponding to fixed objects such as structures and vectors corresponding to the gas. Utilizing this difference between the first vector generated from the original image data and the second vector generated from the processed image data, the second vector is filtered using the first vector, thereby emphasizing and extracting the vectors corresponding to the gas from the second vector. In the extracted vectors, the vectors corresponding to the gas are emphasized, thus becoming vectors with a certain constant length. Therefore, a frequency distribution can be generated from the extracted vectors, and the presence or absence of gas leakage can be determined based on the frequency of the intervals containing the vectors corresponding to the gas in the generated frequency distribution.

[0093] In the technology disclosed in Patent Document 1, since it is based on the premise that gas is present in the captured image, the gas leak monitoring device 10 can accurately indicate whether there is a gas leak even when gas is not present in the captured image. In the gas leak monitoring device 10, during the predetermined image processing emphasizing the gas performed by the image processing unit 13, more than three image data are sometimes used. However, in the processing after the vector generation unit 15, only two original image data and two processed image data are used as image data, eliminating the need to increase the number of image data to improve accuracy as in the technology disclosed in Patent Document 1. Therefore, compared to the method disclosed in Patent Document 1, the gas leak monitoring device 10 of the above embodiment can indicate whether there is a gas leak with high accuracy and in a shorter time.

[0094] (Another structural example of the implementation method)

[0095] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the specific structure is not limited to this embodiment and may include designs that do not depart from the spirit of the present invention.

[0096] (The gas is clearly visible in the original image data)

[0097] When the temperature difference between the fixed object (such as a structure) and the gas is small, the gas is not clearly represented in the raw image data generated by the camera device 2. The gas leak monitoring device 10 described above can filter the vector corresponding to the fixed object to extract the vector corresponding to the gas. Conversely, when the temperature difference between the fixed object and the gas is large, the gas is clearly represented in the raw image data generated by the camera device 2.

[0098] At this time, even without the image processing unit 13 performing the predetermined image processing to emphasize the gas, the vector generation unit 15 includes a vector corresponding to the gas in the first vector generated based on the original image data. Therefore, the gas leak monitoring device 10 may not include the image processing unit 13, the processed image data storage unit 14, and the filtering unit 16. At this time, no processing is performed. Figure 3 The processing of Sb1 to Sb5 involves the vector generation unit 15 performing the following steps when it detects that at least two or more original image data have been recorded in the original image data storage unit 12 by the image data acquisition unit 11. Figure 3The first vector is generated by processing Sb6, and the processing of Sb7, which generates the second vector, is skipped. The processing of Sb8 is also skipped. The frequency distribution generation unit 17 reads the first vector generated by the vector generation unit 15 to replace the vector calculated and output by the filtering unit 16, and performs processing of Sb9 on the read first vector, followed by processing of Sb10 and subsequent processing. The vector generation unit 15 performs processing of Sb11. If the determination is "yes", the processing of Sb6 is performed on the original image data one time later than the original image data of the object processed in the previous Sb6.

[0099] The first vector generated by the vector generation unit 15 includes a vector corresponding to the fixed object and a vector corresponding to the gas. Therefore, even if the first vector is classified according to a predetermined vector length range to generate a frequency distribution with the number of vectors in each range as the frequency, the frequency of this frequency distribution will include the frequency of the vector corresponding to the fixed object. Therefore, when the difference between the length of the vector corresponding to the fixed object and the length of the vector corresponding to the gas is small, it is difficult to determine whether there is a gas leak.

[0100] Furthermore, the moving speed of the fixed object within the image is approximately the same as the moving speed of the camera device 2. Therefore, if the moving speed of the camera device 2 is constant, the range of frequency distribution corresponding to that moving speed can be determined. Thus, the determination unit 18 can, for example, use the frequency of each interval exceeding the length of a preset vector in intervals other than the determined range, and a preset threshold to determine whether there is a gas leak in each image.

[0101] Furthermore, even if the moving speed of the camera device 2 is not constant, as long as there is a significant difference between the moving speed of the camera device 2 and the speed of gas diffusion, the difference between the length of the vector corresponding to the fixed object and the length of the vector corresponding to the gas will increase. Therefore, in the frequency distribution generated by the frequency distribution generation unit 17, there will be a significant difference between the interval indicated by the vector corresponding to the fixed object and the interval indicated by the vector corresponding to the gas. At this time, if the range of the interval indicated by the vector corresponding to the gas can be determined, the determination unit 18 can use the frequency of each interval within the determined range and a preset threshold to determine whether there is gas leakage in each image.

[0102] Therefore, when the gas is clearly presented in the original image data, there is no need for the image processing unit 13 to perform predetermined image processing to emphasize the gas, thus shortening the processing time. In the processing after the vector generation unit 15, only two original image data are used as image data, eliminating the need to increase the number of image data to improve accuracy as in the technology disclosed in Patent Document 1. Therefore, even when the gas is clearly presented in the original image data, compared to the method disclosed in Patent Document 1, it is possible to show whether there is a gas leak with high accuracy and in a shorter time.

[0103] (In the case where a camera device is fixed (Case 1))

[0104] In the above embodiment, for example, the camera device 2 is mounted on a UAV and moved. In contrast, when the camera device 2 is fixed and taken pictures, such as during fixed-point observation, as long as the camera device 2 does not sway due to wind, fixed objects such as structures will not move in the image, so even if optical flow processing is performed, no vector is generated. Therefore, in this case, the gas leak monitoring device 10 may not have a filter unit 16, and the vector generation unit 15 may not generate the first vector. If the second vector generated by the vector generation unit 15 is classified according to a preset vector length range instead of the vector output by the filter unit 16, and the frequency distribution generation unit 17 generates a frequency distribution with the number of vectors in each range as the frequency, then the determination unit 18 can determine whether there is a gas leak in each image based on the frequency distribution. In this case, in the processing after the vector generation unit 15, only two processed image data are used as image data, and it is not necessary to increase the number of image data to improve accuracy as in the technology disclosed in Patent Document 1. Therefore, even with the camera device 2 fixed, it is possible to show whether there is a gas leak with high accuracy and in a shorter time compared to the method disclosed in Patent Document 1.

[0105] (Case with a fixed camera device (Part 2))

[0106] Since a camera device 2 is fixed in place, the temperature difference between the fixed object (such as a structure) and the gas is large, and the gas is clearly presented in the raw image data generated by the camera device 2. In this case, the gas leak monitoring device 10 may not include an image processing unit 13, a processed image data storage unit 14, and a filtering unit 16. The vector generation unit 15 can generate a first vector based on the raw image data stored in the raw image data storage unit 12, but not a second vector. The first vector mainly contains vectors corresponding to the gas. Therefore, if the first vector is classified according to a predetermined vector length range instead of the second vector, and the frequency distribution generation unit 17 generates a frequency distribution using the number of vectors in each range as the frequency, the determination unit 18 can determine whether there is a gas leak in each image based on this frequency distribution.

[0107] When the gas is clearly presented in the original image data, there is no need for the image processing unit 13 to perform predetermined image processing to emphasize the gas, thus shortening the processing time. In the processing after the vector generation unit 15, only two original image data are used as image data, eliminating the need to increase the number of image data to improve accuracy as in the technology disclosed in Patent Document 1. Therefore, even when the camera device 2 is fixed and the gas is clearly presented in the original image data, compared to the method disclosed in Patent Document 1, it is possible to show whether there is a gas leak with high accuracy and in a shorter time.

[0108] (Example of a structure without a decision-making unit)

[0109] In the above embodiment, after the frequency distribution generation unit 17 generates the frequency distribution, the determination unit 18 uses the generated frequency distribution to determine whether there is a gas leak. Alternatively, the determination unit 18 may be omitted, and a person may refer to the frequency of each pre-defined interval (e.g., an interval greater than a pre-defined vector length) in the histogram displayed on the display unit 19 by the frequency distribution generation unit 17 to determine whether there is a gas leak. In this case, the person can also compare the frequency with a threshold to determine whether there is a gas leak, or determine whether there is a gas leak based on the relative magnitude difference of the frequencies in each interval. Therefore, in the latter case where the gas leak is determined based on the relative magnitude difference of the frequencies in each interval, the frequency distribution generation unit 17 may not calculate the frequency of each interval by using the proportion of the number of vectors in each interval as the numerator and the total number of read vectors as the denominator, but instead directly display the number of vectors in each interval as the frequency in the histogram.

[0110] (Other structural examples)

[0111] In the above embodiment, the image data acquisition unit 11 associates the original image data with time information associated with the frame from which the original image data was generated and records this association in the original image data storage unit 12. Alternatively, the image data acquisition unit 11 may not associate the original image data with time information, but instead assign consecutive numbers starting from 1 in a segmented order, for example, sequentially, and record this information in the original image data storage unit 12. In this case, the subsequent processing performed by the image processing unit 13 and the vector generation unit 15 uses the assigned numbers instead of the time information. Specifically, the image processing unit 13... Figure 4 In the Sc2 processing, the original image data used to generate the processed image data is determined according to the assigned number and read from the original image data storage unit 12. The vector generation unit 15... Figure 3 In the processing of Sb6 and Sb7, the original image data and processed image data that are assigned numbers and received from the image processing unit 13 are read out from the original image data storage unit 12 and the processed image data storage unit 14, respectively.

[0112] In the above embodiment, the imaging device 2 captures moving images. Alternatively, the imaging device 2 can also capture still images at different times, associate the still image data with time information indicating the time of capture, and send it to the image data acquisition unit 11. However, the interval between the different times of capture needs to be such that a vector corresponding to an object moving within the image can be generated through optical flow processing; preferably, it is a series of different times, as in the capture of moving images. Alternatively, the imaging device 2 may not associate the still image data with time information, but instead, as described above, associate consecutive numbers starting from 1 with the image data in the order received by the image data acquisition unit 11.

[0113] In the above embodiment, the predetermined image processing of the emphasized gas performed by the image processing unit 13 is not limited to... Figure 4 The image processing shown can also be configured as follows: In the original image data of grayscale, within the range of brightness values ​​from 0 to 255, the brightness values ​​of the gas (e.g., 100 to 150) are set to their original values, and the brightness values ​​of the remaining pixels are set to "0" or "255". Furthermore, image processing that adjusts contrast to emphasize the gas can also be used.

[0114] In the above embodiment, the vector generation unit 15 selects two raw image data that are adjacent in time sequence from the raw image data storage unit 12, and selects two processed image data that are adjacent in time sequence from the processed image data storage unit 14. Alternatively, the vector generation unit 15 may not necessarily select adjacent raw image data. For example, if the raw image data storage unit 12 and the processed image data storage unit 14 store a sufficient number of raw image data, the vector generation unit 15 may, based on selecting one of the raw image data corresponding to the time information output by the image processing unit 13, select another raw image data that is a predetermined time earlier in time sequence than the selected raw image data. In this case, the vector generation unit 15 selects two processed image data from the processed image data storage unit 14 that are associated with the time information of the two selected raw image data. Furthermore, the predetermined time can be preset or changed each time selection is made, but it is preferable to set the predetermined time to a time when the shooting ranges of the two raw image data are not significantly different.

[0115] In the above embodiment, the vector generation unit 15 reads the image processing unit 13... Figure 3 The time information output in the processing of Sb5 is used to select and read the original image data and processed image data corresponding to the read time information in the processing of Sb6 and Sb7. In contrast, during periods when the range captured by the camera device 2 does not change significantly, that is, during periods when the position of a fixed object such as a structure hardly changes, the original image data can be selected without regard to the time information. For example, when the vector generation unit 15 performs one processing of Sb6 and calculates the first vector, it can skip the processing of Sb6 equivalent to the number of processing times during the period when the range captured by the camera device 2 does not change significantly. In the processing of Sb7, when the first vector is output to the filtering unit 16, the most recently generated first vector is output.

[0116] In the above embodiment, the optical flow processing of the Gunnar Farneback method in the flow estimation method is shown as an example of the predetermined image processing of the moving direction and moving speed of the object contained in the image calculated by the vector generation unit 15. However, any optical flow method can be applied, and processing other than optical flow can also be applied.

[0117] In the above embodiment, the filtering unit 16 performs filtering using equation (2). Alternatively, instead of cos(θ / 2) in equation (2), a function other than cos(θ / 2) can be used, which calculates a percentage that is 100% when the angle θ between the second vector and the first vector, which is the same as the position of the pixel, is 0°, and the percentage value approaches 0% as the angle θ approaches 180°.

[0118] In the above embodiment, the gas is set as CO2 gas, but it can also be a gas such as methane or ammonia that absorbs infrared light, or a gas that has absorption wavelengths other than the infrared region. When a gas other than CO2 gas is used as the subject of the photograph, the imaging device 2 is equipped with a filter that allows the absorption wavelengths of the gas to be transmitted.

[0119] In the above embodiment, the image processing unit 13... Figure 4 When generating differential image data in Sc4 processing, if the two original image datasets included in the combination have different shooting ranges, they can be aligned by aligning their shooting ranges before generating the differential image data. Alignment methods can include detecting common computational object points in the two original image datasets and changing their coordinates to make the positions of these common computational object points consistent, or other methods can be applied.

[0120] In the above embodiment, the image data acquisition unit 11 is in Figure 2 In the processing of Sa2, noise removal or grayscale conversion is performed. However, if the image captured by the imaging device 2 has little noise, noise removal may not be performed. Similarly, if the image captured by the imaging device 2 is grayscale instead of color, grayscale conversion may not be performed. Furthermore, the image data acquisition unit 11... Figure 2 In the processing of Sa2, the image data is converted into grayscale image data to generate the original image data by preserving the brightness of the image data. However, the image data can also be converted into grayscale image data to generate the original image data by preserving the lightness of the image data. In this case, in the above embodiment, "brightness" can be replaced with "lightness".

[0121] In the above embodiment, the image processing unit 13... Figure 4 In the Sc5 process, the differential image data is binarized, but binarization may not be performed. At this time, the image processing unit 13 performs the Sc6 process, which takes the differential image data generated in the Sc4 process as the object, that is, it performs the process of accumulating the brightness value of each pixel of the differential image data generated in the Sc4 process to generate processed image data, and then performs the Sc7 and Sc8 processes.

[0122] In the above embodiment, the filtering unit 16 excludes vectors from the outer edge of the image. However, if the reliability of vectors from the outer edge of the image is high, vectors from the outer edge of the image may not be excluded.

[0123] In the above embodiments, such as Figure 3 As shown in the processing of Sb6 and Sb7, the vector generation unit 15 generates the first vector and then generates the second vector. However, conversely, the first vector can be generated after the second vector is generated, or the processing of generating the first vector and generating the second vector can be performed in parallel.

[0124] In the above embodiment, the determination unit 18 uses the frequency of each interval above the length of a preset vector and a preset threshold to determine whether there is a gas leak in each image. In contrast, when the length of the vector corresponding to the gas is empirically known, it is possible not to set an interval above the length of the preset vector, but to preset an interval including the length of the vector corresponding to the gas, and use the frequency of each preset interval and a preset threshold to determine whether there is a gas leak in each image.

[0125] In the above embodiment, the image processing unit 13 receives an instruction signal indicating continued processing from the determination unit 18 and performs processing accordingly. Figure 3 The processing of Sb11. Alternatively, the determination unit 18 may not output an instruction signal indicating continued processing, and the image processing unit 13 may perform the processing of Sb11 after processing Sb5.

[0126] In the above embodiments, Figure 4 Sc5 processing and Figure 3 In the processing of Sb10, the image processing unit 13 and the determination unit 18 perform threshold determination processing to determine whether the value of the object is above the threshold. In contrast, in the processing of Sc5 and Sb10, depending on the method for determining the threshold, the processing can also be set to determine whether the value of the object exceeds the threshold.

[0127] (Computer architecture)

[0128] Figure 12This is a schematic block diagram illustrating the structure of a computer according to at least one embodiment. The computer 90 includes a processor 91, a main memory 92, a storage device 93, and an interface 94. The aforementioned gas leak monitoring device 10 is installed in the computer 90. Furthermore, the operations of each of the aforementioned processing units—namely, the image data acquisition unit 11, the image processing unit 13, the vector generation unit 15, the filtering unit 16, the frequency distribution generation unit 17, and the determination unit 18—are stored in the storage device 93 in the form of a program. The processor 91 reads the program from the storage device 93, expands it to the main memory 92, and executes the aforementioned processing according to the program. Furthermore, the processor 91 secures, according to the program, storage areas in the main memory 92 or the storage device 93 corresponding to the aforementioned original image data storage unit 12 and processed image data storage unit 14. The display unit 19 is connected via the interface 94. Therefore, the display unit 19 may or may not be a component of the computer 90, i.e., a component of the gas leak monitoring device 10 as described above.

[0129] The program can also be used to implement a portion of the functions that enable the computer 90 to perform. For example, the program can perform functions by combining with other programs already stored in storage device 93 or with other programs installed in other devices. Additionally, in other embodiments, the computer may possess a custom LSI (Large Scale Integrated Circuit) such as a PLD (Programmable Logic Device) or alternative to the above-described structure. Examples of PLDs include PAL (Programmable Array Logic), GAL (Generic Array Logic), CPLD (Complex Programmable Logic Device), and FPGA (Field Programmable Gate Array). In this case, some or all of the functions implemented by the processor can be implemented by this integrated circuit.

[0130] Examples of storage devices 93 include HDDs (Hard Disk Drives), SSDs (Solid State Drives), magnetic disks, optical disks, CD-ROMs (Compact Disc Read Only Memory), DVD-ROMs (Digital Versatile Disc Read Only Memory), and semiconductor memory. Storage device 93 can be an internal medium directly connected to the bus of computer 90, or an external medium connected to computer 90 via interface 94 or a communication line. Furthermore, when the program is distributed to computer 90 via the communication line, computer 90 receiving the distribution can expand the program into main memory 92 to execute the aforementioned processing. In at least one embodiment, storage device 93 is a non-transitory tangible storage medium.

[0131] <Postscript>

[0132] The gas leak monitoring device 10 described in the embodiments of the present invention can be understood, for example, as follows.

[0133] (1) The gas leak monitoring device 10 according to the first method includes: an image data acquisition unit 11, which acquires image data generated by taking pictures at different times as raw image data; a vector generation unit 15, which selects two raw image data in time sequence from the raw image data, and performs predetermined image processing on the selected two raw image data to calculate the movement direction and movement speed of the objects contained in the images, thereby generating a vector representing the movement direction and movement speed for each pixel contained in the raw image data; and a frequency distribution generation unit 17, which classifies the vectors generated by the vector generation unit according to a predetermined vector length range to generate a frequency distribution for gas leak monitoring. According to this method and the following methods, it is possible to indicate whether there is a gas leak with high accuracy and in a shorter time.

[0134] (2) The gas leak monitoring device 10 involved in the second method is the gas leak monitoring device 10 of (1), which includes an image processing unit 13. The image processing unit 13 performs predetermined image processing on the original image data to emphasize the gas and generates processed image data. The vector generation unit performs the following processing: using the processed image data to replace the original image data, and generating a vector representing the direction of movement and the speed of movement for each pixel contained in the processed image data. According to this method, even if the gas is not clearly presented in the original image data, a vector corresponding to the gas can be generated based on the processed image data that emphasizes the gas.

[0135] (3) The gas leak monitoring device 10 involved in the third method is the gas leak monitoring device 10 of (1), which includes: an image processing unit 13, which performs predetermined image processing to emphasize the gas in the original image data to generate processed image data; and a filtering unit 16, which emphasizes and extracts vectors corresponding to the gas. The vector generation unit performs the following processing: the vector generated for each pixel contained in the original image data is set as the first vector, and then two processed image data are selected from the processed image data in time sequence, and the selected two processed image data are subjected to predetermined image processing to calculate the moving direction and moving speed of the object contained in the image, thereby generating a vector representing the moving direction and moving speed for each pixel contained in the processed image data and setting it as the second vector. The filtering unit performs the following processing: the second vector of each pixel is filtered using the first vector that is consistent with the pixel and the second vector. The frequency distribution generation unit performs the following processing: the vector extracted by the filtering unit is used to replace the vector generated by the vector generation unit to generate the frequency distribution for gas leak monitoring. According to this method, even if the camera device taking the picture is moving and the gas is not clearly presented in the original image data, processed image data that emphasizes the gas is generated. Based on the difference between the first vector generated from the original image data and the second vector generated from the processed image data, filtering is performed, thereby enabling the emphasis and extraction of the vector corresponding to the gas. The extracted vector is used to indicate whether there is a gas leak.

[0136] (4) The gas leak monitoring device 10 involved in the fourth method is the gas leak monitoring device 10 of (3), wherein the vector generation unit performs the following processing: when selecting two original image data, it selects two original image data that are adjacent in time sequence; when selecting two processed image data, it selects processed image data corresponding to each of the two selected original image data. According to this method, the two original image data of the generation source of the first vector become adjacent in time sequence, and the two processed image data of the generation source of the second vector become processed image data that correspond to each of the two original image data and are adjacent in time sequence. Thus, by using image data that are adjacent in time sequence, the accuracy of the first vector and the second vector is improved, and by using processed image data corresponding to each of the two original image data, the shooting range is consistent in the original image data and the processed image data, so the vector corresponding to the gas can be extracted with higher accuracy.

[0137] (5) The gas leak monitoring device 10 involved in the fifth method is the gas leak monitoring device 10 of (3) or (4), wherein the filtering unit performs filtering as follows: adding the second vector to a vector having a length obtained by multiplying the length of the first vector by a ratio and having a direction opposite to the second vector, wherein the ratio is 100% when the angle between the second vector and the first vector whose pixel position is the same as the second vector is 0°, and the percentage value approaches 0% as the angle approaches 180°. According to this method, the first vector can be used to exclude vectors corresponding to fixed objects such as structures from the second vector, and vectors corresponding to gases are emphasized.

[0138] (6) The gas leak monitoring device 10 involved in the sixth method is any one of (2) to (5) gas leak monitoring devices 10, wherein the image processing unit performs the following image processing as a predetermined image processing to emphasize the gas: selecting a combination of two original image data arranged sequentially in time sequence from the original image data of the processing object and a plurality of original image data acquired by the image data acquisition unit at a time earlier than the original image data, generating differential image data for each of the selected combinations, accumulating the pixel value of the generated differential image data for each pixel, thereby generating the processed image data corresponding to the original image data of the processing object. According to this method, even if the gas is not clearly presented in the original image data, processed image data emphasizing the gas can be generated.

[0139] (7) The gas leak monitoring device 10 involved in the seventh method is any one of (1) to (6) gas leak monitoring devices 10, which includes a determination unit 18 for determining whether there is a gas leak. The frequency distribution generation unit performs the following processing: generating a frequency distribution in which the frequency of each interval is the proportion of the number of vectors in each interval as the numerator and the total number of vectors as the denominator. The determination unit performs the following processing: determining whether there is a gas leak based on the frequency of the interval and a preset threshold. According to this method, there is no need for a person to refer to a histogram to determine whether there is a gas leak. The determination unit 18 can determine whether there is a gas leak based on the frequency distribution data generated by the frequency distribution generation unit 17.

[0140] (8) The gas leak monitoring device 10 involved in the eighth method is any one of (1) to (7) gas leak monitoring devices 10, wherein the filtering unit performs the following processing: excluding the vectors corresponding to the pixels at the outer edge of the image data from the vectors after emphasizing and extracting the vectors corresponding to the gas. According to this method, the vectors used in generating the frequency distribution can be reduced to vectors with high reliability, thereby improving the accuracy of the frequency distribution, and consequently enabling high-precision differentiation of whether there is a gas leak.

[0141] Industrial availability

[0142] The gas leak monitoring device, gas leak monitoring method and procedure according to the present invention can indicate whether there is a gas leak with high accuracy and in a shorter time.

[0143] Symbol Explanation

[0144] 1-Gas leak monitoring system, 2-Camera device, 3-Communication network, 10-Gas leak monitoring device, 11-Image data acquisition unit, 12-Raw image data storage unit, 13-Image processing unit, 14-Processed image data storage unit, 15-Vector generation unit, 16-Filtering unit, 17-Frequency distribution generation unit, 18-Judgment unit, 19-Display unit.

Claims

1. A gas leak monitoring device, comprising: The image data acquisition unit acquires image data generated by taking pictures at different times as raw image data; The vector generation unit selects two original image data points from the original image data in chronological order, and performs predetermined image processing on the selected two original image data points to calculate the direction and speed of movement of objects contained in the images, thereby generating a vector representing the direction and speed of movement for each pixel contained in the original image data; and The frequency distribution generation unit classifies the vectors generated by the vector generation unit according to a pre-set vector length range to generate a frequency distribution for gas leak monitoring.

2. The gas leak monitoring device according to claim 1, further comprising an image processing unit, wherein the image processing unit performs predetermined image processing on the original image data to emphasize the gas and generate processed image data. The vector generation unit performs the following processing: The processed image data is used to replace the original image data, and a vector representing the direction and speed of movement is generated for each pixel contained in the processed image data.

3. The gas leak monitoring device according to claim 1, comprising: The image processing unit performs predetermined image processing on the original image data to emphasize the gas, thereby generating processed image data; and The filter section emphasizes and extracts the vector corresponding to the gas. The vector generation unit performs the following processing: The vector generated for each pixel in the original image data is designated as the first vector. Then, two processed image data are selected sequentially from the processed image data, and predetermined image processing is performed on the selected two processed image data to calculate the direction and speed of movement of the objects contained in the images. This generates a vector representing the direction and speed of movement for each pixel in the processed image data and designates it as the second vector. The filtering section performs the following processing: For each pixel, the second vector is filtered using the first vector that is consistent with the pixel's second vector. The frequency distribution generation unit performs the following processing: The frequency distribution for gas leak monitoring is generated by replacing the vector generated by the vector generation unit with the vector extracted by the filtering unit.

4. The gas leak monitoring device according to claim 3, wherein, The vector generation unit performs the following processing: When selecting two original image data, two original image data that are adjacent in time sequence are selected. When selecting two processed image data, processed image data corresponding to each of the two selected original image data are selected.

5. The gas leak monitoring device according to claim 3, wherein, The filtering unit performs the following filtering: The vector with a length obtained by multiplying the length of the first vector by a ratio and having a direction opposite to the second vector is added to the second vector. The ratio is 100% when the angle between the second vector and the first vector, whose pixel position is the same as the second vector, is 0°, and the percentage value approaches 0% as the angle approaches 180°.

6. The gas leak monitoring device according to any one of claims 2 to 5, wherein, The image processing unit performs the following image processing as a predetermined image processing to emphasize the gas: From the original image data of the object being processed and a plurality of original image data acquired by the image data acquisition unit at an earlier time than the original image data, a combination of two original image data arranged sequentially in time series is selected, differential image data for each of the selected combinations is generated, and the pixel value of the generated differential image data is accumulated for each pixel, thereby generating the processed image data corresponding to the original image data of the object being processed.

7. The gas leak monitoring device according to any one of claims 1 to 5, comprising a determination unit for determining whether a gas leak exists. The frequency distribution generation unit performs the following processing: Generate the frequency distribution with the proportion of the number of vectors in each interval as the numerator and the total number of vectors as the denominator, which is the frequency of each interval. The determination unit performs the following processing: The presence or absence of gas leakage is determined based on the frequency of the preset interval and the preset threshold.

8. The gas leak monitoring device according to any one of claims 3 to 5, wherein, The filtering section performs the following processing: Vectors corresponding to pixels at the outer edge of the image data are excluded from the vectors after highlighting and extracting the vectors corresponding to the gas.

9. A method for monitoring gas leaks, comprising the following steps: Acquire image data generated by taking pictures at different times as raw image data; Two original image data points are selected sequentially from the acquired original image data in time series order. Predetermined image processing is then performed on the selected two original image data points to calculate the direction and speed of movement of objects contained within the images. This generates a vector representing the direction and speed of movement for each pixel contained in the original image data. The generated vectors are classified according to a pre-defined vector length range to generate a frequency distribution for gas leak monitoring.

10. A program that causes a computer to perform the following steps: Acquire image data generated by taking pictures at different times as raw image data; Two original image data points are selected sequentially from the acquired original image data in time series order. Predetermined image processing is then performed on the selected two original image data points to calculate the direction and speed of movement of objects contained within the images. This generates a vector representing the direction and speed of movement for each pixel contained in the original image data. The generated vectors are classified according to a pre-defined vector length range to generate a frequency distribution for gas leak monitoring.

Images