Array grating-based oil tank surface temperature field image fire monitoring method

Abstract

The application discloses a kind of oil tank surface temperature field image fire monitoring method based on array grating, temperature signal of entire detection area is collected by array grating temperature sensing optical cable laid on oil tank surface, records full section temperature signal matrix in each time period, draws temperature field image, and temperature field image is handled as foreground and background part, reference temperature sequence is obtained by processing temperature field image background, it is found whether there is the area reaching high temperature threshold in space-time temperature field image, if there is, high temperature alarm is sent out;Whether exceeding singular region is judged by setting temperature rise threshold, and temperature rise alarm is carried out.The monitoring method effectively suppresses the interference caused by the temperature difference introduced by the laying position of temperature sensing optical cable, environment difference, avoids the misjudgment of temperature rise alarm caused by the influence of factors such as solar radiation and oil properties on the surface temperature of the oil tank, and makes the monitoring of abnormal changes in the surface temperature of the oil tank more accurate.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of array grating temperature sensing, in particular to an oil tank surface temperature field image fire monitoring method based on array grating. BACKGROUND

[0002] Temperature fire detection is a method of early detection and prediction of oil tank fire, which monitors the temperature of the oil tank surface in real time and issues an alarm when the temperature changes abnormally or reaches the fire alarm threshold. Array grating temperature sensor has the advantages of electrical insulation, anti-electromagnetic interference, corrosion resistance, high sensitivity, high temperature measurement accuracy, long service life, easy networking, etc., and is very suitable for oil tank fire monitoring. The alarm method used by the grating temperature oil tank fire alarm system is temperature rise alarm and high temperature alarm. When the system monitors that the temperature of a certain detection position in the previous period rises by more than the temperature rise threshold, a temperature rise alarm signal is issued; when the temperature value of a certain detection position exceeds the high temperature threshold, a high temperature alarm signal is issued.

[0003] However, this alarm method does not fully utilize the distributed measurement advantages of array grating sensing technology to analyze the spatial information of the temperature field. That is, when the temperature of the oil tank surface is affected by factors such as solar radiation and oil properties, false alarms of temperature rise alarms are likely to occur. In addition, due to the different environments of the temperature sensing cable laying position, the reference temperature will also show a slow change trend in space, so it is not appropriate to use a uniform threshold for high temperature alarm. SUMMARY

[0004] The present application provides an oil tank surface temperature field image fire monitoring method based on array grating, which is simple and effective, effectively suppresses the interference caused by temperature differences introduced by the laying position and environment of the temperature sensing cable, avoids false alarms of temperature rise alarms caused by the influence of factors such as solar radiation and oil properties on the temperature of the oil tank surface, and makes the monitoring of abnormal changes in the temperature of the oil tank surface more accurate.

[0005] To achieve the above-mentioned purpose, the present application provides an oil tank surface temperature field image fire monitoring method based on array grating, which specifically comprises the following steps:

[0006] Step 1, laying array grating temperature sensing cable as temperature sensor on the surface of the oil tank, each detection unit in the array grating temperature sensing cable is used to detect the temperature signal of the entire region;

[0007] Splice the data collected at each moment to obtain a full section temperature signal matrix within t seconds, draw a temperature field image according to the full section temperature signal matrix, the matrix horizontal axis direction of the temperature field image represents spatial distance, and the matrix vertical axis direction represents time, forming a space-time temperature field image and the corresponding matrix;

[0008] Step 2, processing the temperature field image into foreground and background parts;

[0009] Step 3, taking the background part of the temperature field image as the reference temperature of this time, calculating the mean value of each column to obtain the reference temperature sequence;

[0010] Step 4, taking the reference temperature sequence plus a set value Thigh as the high temperature threshold of this time, searching whether there is a region reaching the high temperature threshold in the space-time temperature field image, if there is, counting the spatial coverage of this region, and issuing a high temperature alarm;

[0011] Step 5, performing Blob detection on the foreground part of the temperature field image, taking the reference temperature sequence plus a set value Trise as the temperature rise threshold of this time, for the singular region obtained by Blob detection, if there is an abnormal temperature rise region whose size exceeds the threshold TminArea and temperature exceeds the temperature rise threshold, counting the spatial coverage of this region, and issuing a temperature rise alarm.

[0012] Further, in step 2, the temperature field image is first subjected to top-hat transformation processing, and then the global threshold is extracted by OTSU algorithm, and is divided into foreground and background parts.

[0013] Further, the OTSU algorithm step is specifically:

[0014] Suppose there is a threshold T H , T H is the threshold for dividing the foreground and background of the temperature field image, all pixels of the temperature field image are divided into two categories C1 and C2, where C1 is less than T H , C2 is greater than T H , and the mean values of the two categories of pixels are m1 and m1 respectively, the global mean value of the image is m G , and the probabilities of the pixels being divided into C1 and C2 categories are p1 and p2 respectively, so:

[0015] p1*m1+p2*m2=m G (1)

[0016] p1+p2=1 (2)

[0017] According to the concept of variance, the inter-class variance expression is:

[0018] σ 2 =p1(m1-m G ) 2 +p2(m2-m G ) 2 (3)

[0019] Simplify the above formula and bring formula (1) into formula (3), we can get:

[0020] σ 2 =p1p2(m1-m2)2 (4)

[0021] The maximum value L in the temperature field image is obtained statistically, where:

[0022]

[0023]

[0024]

[0025] According to formula (4), traverse the values ​​0-L to find the k that maximizes formula (4), k∈(0,L), which is the global threshold.

[0026] Furthermore, in step 3, the search is conducted to determine if there are regions in the spatial-temporal temperature field image that reach the high-temperature threshold. The specific method is as follows:

[0027] Let n be the number of units in the detected temperature signal. Then, the high-temperature threshold is a 1xn sequence. Let t be the number of rows in the spatial-temporal temperature field matrix, which is the spatial-temporal temperature field matrix formed within time t. The high-temperature threshold sequence is copied and stacked into a txn matrix, making its size the same as the spatial-temporal temperature field matrix. Then, the difference matrix is ​​obtained by subtracting the spatial-temporal temperature field matrix from the high-temperature threshold matrix. We then check if there are regions >0 in this difference matrix. If so, we calculate the spatial coverage of this region and issue a high-temperature alarm.

[0028] Furthermore, the arrayed grating temperature-sensing optical cable described in step 1 contains at least 270 detection units and has a measurement length of 1 km.

[0029] In step 2, the sampling rate for acquiring the detected temperature signal is 1Hz.

[0030] Furthermore, in step 5, the Laplacian of Gaussian algorithm is used to perform Blob detection on the foreground portion of the temperature field image.

[0031] Compared with existing technologies, this method for fire monitoring of oil tank surface temperature field images based on array gratings collects temperature signals of the entire detection area through array grating temperature sensing optical cables laid on the surface of the oil tank. It records the temperature signal matrix of the entire segment in various time periods, draws temperature field images, forms a spatial-temporal temperature field image and corresponding matrix, and processes the temperature field images to obtain a reference temperature sequence. Therefore, by extracting the gradual change trend of the reference temperature sequence in space, it effectively suppresses the interference caused by temperature differences introduced by different laying positions of temperature sensing optical cables and different environments, and avoids false judgments of temperature rise alarms caused by factors such as solar radiation and oil properties affecting the surface temperature of the oil tank.

[0032] Since the singular region of the temperature field image is detected as the basis for temperature rise alarm, the spatial distribution and time distribution of the temperature field are combined and analyzed by using the distributed measurement advantage of the array grating, so that the abnormal change of the temperature on the surface of the oil tank is more accurately monitored. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application is further described in detail. It should be understood that the specific embodiments described herein are only used to explain the present application and not to limit the present application.

[0034] The oil tank surface temperature field image fire monitoring system based on array grating specifically comprises a data acquisition module, a temperature field image analysis module, a reference temperature calculation module, a high temperature alarm module and a temperature rise alarm module.

[0035] The data acquisition module comprises an array grating temperature sensing optical cable laid on the surface of the oil tank for temperature sensing, which is used to detect the temperature signal of the entire region and draw a temperature field image according to the full section temperature signal matrix. In the temperature field image, the matrix horizontal axis direction represents the spatial distance, and the matrix vertical axis direction represents the time, that is, a space-time temperature field image and a corresponding matrix are formed.

[0036] The temperature field image analysis module analyzes and processes the temperature field image into foreground and background parts.

[0037] The reference temperature calculation module calculates the reference temperature sequence according to the background part of the temperature field image.

[0038] The high temperature alarm module sets a high temperature threshold according to the reference temperature sequence, finds whether there is a region reaching the high temperature threshold in the space-time temperature field, if there is, counts the spatial coverage range of the region, and issues a high temperature alarm.

[0039] The temperature rise alarm module sets a temperature rise threshold according to the reference temperature sequence. If there is a singular region whose size exceeds the threshold and the temperature exceeds the temperature rise threshold, the spatial coverage range of the region is counted, and a temperature rise alarm is issued. The singular region is obtained by Blob detection on the foreground part of the temperature field image.

[0040] The oil tank surface temperature field image fire monitoring method based on array grating specifically comprises the following steps:

[0041] Step 1, the array grating temperature sensing optical cable is used as a temperature sensor and laid on the surface of the oil tank. Each detection unit in the array grating temperature sensing optical cable is used to detect the temperature signal of the entire region. For example, the array grating temperature sensing optical cable can contain at least 270 detection units, the measurement length is about 1km, and it is laid on the surface of the oil tank.

[0042] The detected temperature signal is collected, and the preferred sampling rate is 1 Hz;

[0043] The data collected at each time is spliced to obtain a full-segment temperature signal matrix within t seconds, such as a full-segment temperature signal matrix within the first 600 seconds, and a temperature field image is drawn according to the full-segment temperature signal matrix, wherein the matrix horizontal axis direction represents spatial distance, and the matrix vertical axis direction represents time, i.e., a space-time temperature field image and a corresponding matrix are formed;

[0044] Step 2, the temperature field image is processed and divided into foreground and background parts;

[0045] Step 3, the background part of the temperature field image is taken as the reference temperature of this time, the mean value of each column is calculated, and a reference temperature sequence is obtained;

[0046] Step 4, the reference temperature sequence is added to a set value Thigh as a high temperature threshold value at this time, and it is checked whether there is a region reaching the high temperature threshold value in the space-time temperature field image, if there is, the spatial coverage range of the region is counted, and a high temperature alarm is issued;

[0047] Step 5, the Laplacian of Gaussian algorithm is used to detect the Blob of the foreground part of the temperature field image, the set value Trise is added to the reference temperature sequence as a temperature rise threshold value at this time, for the singular region obtained by Blob detection, if there is a singular region with a size exceeding an abnormal temperature rise region threshold value TminArea, preferably 20 pixel points, and the temperature exceeds the temperature rise threshold value, the spatial coverage range of the region is counted, and a temperature rise alarm is issued.

[0048] The set value Thigh, the set value Trise, and the abnormal temperature rise region threshold value TminArea are parameters set according to the use environment, conditions, etc., such as setting the set value Thigh = 30℃, the set value Trise = 10℃, and ensuring that the set value Trise < the set value Thigh when setting;

[0049] Further, in step 2, the temperature field image is first processed by top-hat transformation, and then the global threshold value is extracted by OTSU algorithm to divide the temperature field image into foreground and background parts;

[0050] Further, the OTSU algorithm step is specifically as follows:

[0051] Suppose there is a threshold value T H , T H is a temperature field image foreground and background division threshold value, which is part of the OTSU algorithm, and all pixels of the temperature field image are divided into two categories C1 and C2, wherein C1 is less than T H , C2 is greater than T HThen the mean of the two classes of pixels is m1, m1 respectively, and the global mean of the image is m G The probability of the pixels being classified into class C1 and C2 is p1, p2 respectively. Therefore, we have:

[0052] p1*m1+p2*m2=m G (1)

[0053] p1+p2=1 (2)

[0054] According to the concept of variance, the inter-class variance is expressed as:

[0055] σ 2 = p1(m1-m G ) 2 + p2(m2-m G ) 2 (3)

[0056] Simplifying the above formula and substituting formula (1) into formula (3), we have:

[0057] σ 2 = p1p2(m1-m2) 2 (4)

[0058] The maximum value L in the temperature field image is obtained by statistics, and the value k (k∈(0, L)) that maximizes formula (4) is the OTSU threshold value, where

[0059]

[0060]

[0061]

[0062] According to formula (4), the value of k that maximizes formula (4) is obtained by traversing the values from 0 to L, which is the global threshold value.

[0063] Further, in step 4, it is determined whether there is a region reaching the high temperature threshold value in the space-time temperature field image, and the specific method is as follows:

[0064] Let the number of units of the detected temperature signal be n, then the high temperature threshold value is a sequence of 1 row and n columns, and let the number of rows of the space-time temperature field matrix be t, that is, the space-time temperature field matrix formed in t time. The high temperature threshold value sequence is copied and stacked into a matrix of t rows and n columns, so that the matrix size is the same as the space-time temperature field matrix. Then, the difference matrix is obtained by subtracting the space-time temperature field matrix from the high temperature threshold matrix, and it is determined whether there is a region with a value greater than 0 in the difference matrix. If there is, the spatial coverage of the region is counted, and a high temperature alarm is issued.

[0065] The array grating-based oil tank surface temperature field image fire monitoring method uses the array grating temperature sensing optical cable laid on the surface of the oil tank to collect temperature signals of the entire detection area, records the full section temperature signal matrix in each time period, draws a temperature field image, forms a space-time temperature field image and a corresponding matrix, and processes the temperature field image to obtain a reference temperature sequence, so that the slow change trend of the reference temperature sequence in space is extracted, the interference caused by temperature differences introduced by the laying position of the temperature sensing optical cable and the environment is effectively suppressed, and the influence of factors such as solar radiation and oil properties on the temperature of the surface of the oil tank is avoided to prevent false alarms caused by temperature rise.

[0066] Since the temperature field image is detected as the basis for temperature rise alarm, the spatial distribution and time distribution of the temperature field are combined and analyzed by using the distributed measurement advantage of the array grating, and the temperature abnormal change of the surface of the oil tank is more accurately monitored.

Claims

1. An array grating-based oil tank surface temperature field image fire monitoring method, specifically comprising the following steps: Step 1: An array grating temperature sensing optical cable is laid on the surface of an oil tank as a temperature sensor, and each detection unit in the array grating temperature sensing optical cable is used to detect the temperature signal of the entire region; The data collected at each moment is spliced to obtain a full-segment temperature signal matrix within t seconds, and a temperature field image is drawn according to the full-segment temperature signal matrix, wherein the matrix horizontal axis direction represents spatial distance, and the matrix vertical axis direction represents time, forming a space-time temperature field image and a corresponding matrix; Step 2: The temperature field image is first subjected to a top-hat transformation process, and then an OTSU algorithm is used to extract a global threshold value, which is divided into foreground and background parts; Step 3: The background part of the temperature field image is taken as the reference temperature of this time period, the mean value of each column is calculated, and a reference temperature sequence is obtained; Step 4: The reference temperature sequence is added with a set value Thigh as the high temperature threshold value of this time, and it is checked whether there is a region reaching the high temperature threshold value in the space-time temperature field image, if there is, the spatial coverage range of the region is counted, and a high temperature alarm is issued; Step 5: The foreground part of the temperature field image is subjected to Blob detection, the reference temperature sequence is added with a set value Trise as the temperature rise threshold value of this time, and for the abnormal temperature rise region threshold value TminArea and the singular region with a temperature exceeding the temperature rise threshold value obtained by Blob detection, if there is a region with a size exceeding the threshold value, the spatial coverage range of the region is counted, and a temperature rise alarm is issued.

2. The arrayed-grating-based tank surface temperature field image fire monitoring method according to claim 1, characterized in that, The OTSU algorithm steps are specifically as follows: assuming that there is a threshold TH, TH is a temperature field image foreground and background division threshold, all pixels of the temperature field image are divided into two categories C1 and C2, wherein C1 is less than T H , C2 is greater than T H , the average values of the two categories of pixels are m1 and m2 respectively, the global average value of the image is m G , and the probabilities of the pixels being divided into the C1 and C2 categories are p1 and p2 respectively, so that: p1 * m1+ p2 * m2= m G (1) p1+ p2=1 (2) According to the concept of variance, the inter-class variance expression is: σ 2 = p1(m1- m G ) 2 + p2(m2- m G ) 2 (3) Simplify the above formula and bring formula (1) into formula (3) to obtain: σ2 = p1 p2 (m1 - m2) 2 (4) The maximum value L in the temperature field image is obtained by statistics, wherein: （5） （6） （7） According to formula (4), the value 0-L is traversed, the maximum k of formula (4) is calculated, k∈(0, L), and the global threshold value is obtained.

3. The arrayed-grating-based oil tank surface temperature field image fire monitoring method according to claim 1 or 2, characterized in that, The specific method for checking whether there is a region reaching the high temperature threshold value in the space-time temperature field image in step 3 is as follows: Let the number of unit detection temperature signals be n, then the high temperature threshold value is a 1-row n-column sequence, let the number of rows of the space-time temperature field matrix be t, that is, the space-time temperature field matrix formed within t time, copy and stack the high temperature threshold value sequence into a t-row n-column matrix, so that the matrix size is the same as that of the space-time temperature field matrix, then subtract the space-time temperature field matrix from the high temperature threshold value matrix to obtain a difference matrix, and check whether there is a region >0 in the difference matrix; if there is, count the spatial coverage range of the region, and issue a high temperature alarm.

4. The arrayed-grating-based tank surface temperature field image fire monitoring method according to claim 3, characterized in that, The array grating temperature sensing optical cable in step 1 contains at least 270 detection units, and the measurement length is 1 km; the sampling rate of the collected temperature signal in step 2 is 1 Hz.

5. The arrayed-grating-based tank surface temperature field image fire monitoring method according to claim 3, characterized in that, In step 5, the Laplacian of Gaussian algorithm is used to perform Blob detection on the foreground part of the temperature field image.

Images