Sulfur hexafluoride ring main unit gas leakage detection method and system based on thermal imaging

Abstract

The invention relates to the technical field of image processing, and discloses a sulfur hexafluoride ring main unit gas leakage detection method and system based on thermal imaging, and the method comprises the steps: obtaining an infrared thermal imaging multispectral sequence image of a ring main unit; identifying a plurality of candidate regions of the multispectral sequence image, selecting a gas reference region, and calculating an average multispectral vector of the gas reference region as a gas reference spectrum; calculating the intra-group variance of each spectrum channel, and determining a weight vector and an angle threshold value; calculating a weighted spectral angle between the multispectral vector of the pixel point to be detected and the gas reference spectrum; judging the to-be-detected pixel point which simultaneously meets the following conditions as a gas leakage point: the weighted spectral angle is smaller than an angle threshold value, and the temperature of the to-be-detected pixel point is lower than a preset temperature threshold value; and generating a visual image of the gas leakage area according to all the judged gas leakage points. According to the invention, the anti-interference capability of detection is improved, and reliable identification of sulfur hexafluoride gas leakage can be realized under a complex background.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, in particular to a sulfur hexafluoride ring network cabinet gas leakage detection method and system based on thermal imaging. BACKGROUND

[0002] Sulfur hexafluoride gas is used as a key medium in high-voltage electrical equipment such as ring network cabinets and gas-insulated switches due to its excellent insulation and arc-extinguishing performance. The leakage of sulfur hexafluoride will cause double hidden dangers. On the one hand, it will weaken the insulation performance of the equipment, directly threatening the safe and stable operation of the power grid; on the other hand, as a very strong greenhouse gas, its greenhouse effect is as high as 23500 times that of carbon dioxide, and it is extremely difficult to decompose in the atmosphere, once leaked, it will pose a long-term and serious threat to the environment.

[0003] At present, traditional sulfur hexafluoride gas leakage detection methods, such as wrapping leak detection and handheld gas leak detector, mainly rely on contact or close-range measurement. These detection methods not only have low detection efficiency and cannot meet the needs of large-scale and long-distance inspection, but also their detection results are easily disturbed by environmental factors such as wind speed. More importantly, when working in some high-voltage dangerous areas, the personal safety of the detection personnel also faces high risks.

[0004] In view of the above defects, infrared thermal imaging technology provides a non-contact remote leakage detection method, which is mainly based on two physical properties of sulfur hexafluoride: first, when the gas leaks from the high-voltage equipment, it will rapidly cool down due to the adiabatic expansion effect, forming a low-temperature gas cloud; second, this gas cloud can strongly absorb infrared radiation of a specific waveband. The two characteristics work together to make the leaked gas cloud show a significant difference in temperature and spectrum from the surrounding background in the infrared image, thereby realizing visual detection.

[0005] However, infrared detection technology still faces challenges in practical application. Complex on-site environment can easily lead to false positives, for example, radiation differences of different materials on the surface of the equipment, changes in background temperature, and even reflections of sunlight or high-temperature objects, which can all form low-temperature or low-radiation pseudo-images similar to gas leakage on the image. In addition, the form, concentration and temperature of the gas will dynamically change with factors such as leakage rate, environmental temperature and humidity, and wind power, making its thermal and spectral characteristics in the image unstable, increasing the difficulty of accurate identification, and easily producing false positives. SUMMARY

[0006] The present application provides a sulfur hexafluoride ring network cabinet gas leakage detection method and system based on thermal imaging to solve the problem of low accuracy of leakage detection results and easy false positives in the prior art.

[0007] In a first aspect, the present application provides a sulfur hexafluoride ring network cabinet gas leakage detection method based on thermal imaging, comprising the following steps: S1, acquiring an infrared thermal imaging multispectral sequence image of the ring main unit; identifying a plurality of candidate regions with low temperature and uniform texture in each frame image of the multispectral sequence image through mean shift clustering; selecting a region with the lowest average temperature from the candidate regions as a gas reference region, and calculating an average multispectral vector of the gas reference region as a gas reference spectrum; S2, calculating the within-group variance of each spectral channel in the gas reference region, and determining a weight vector and an angle threshold based on the within-group variance; for any to-be-detected pixel point in the frame image, calculating the weighted spectral angle between the multispectral vector of the to-be-detected pixel point and the gas reference spectrum based on the weight vector; S3, determining the to-be-detected pixel point as a gas leakage point if the following conditions are met simultaneously: the weighted spectral angle is less than the angle threshold, and the temperature of the to-be-detected pixel point is lower than a preset temperature threshold; generating a visual image of a gas leakage region according to all determined gas leakage points.

[0008] Preferably, the identification of the plurality of candidate regions with low temperature and uniform texture in each frame image of the multispectral sequence image through mean shift clustering comprises: S11: initializing a cluster center, and taking all pixel points in the frame image as an initial sample point set; S12: setting a spatial domain bandwidth to 15, and a color domain bandwidth to 20; S13: calculating a mean shift vector of each sample point within the bandwidth range, and moving the sample point to the terminal point of the mean shift vector; S14: repeating step S13 until the moving distance of all sample points is less than a preset convergence threshold, and completing clustering; S15: calculating the average temperature and the energy value of the gray level co-occurrence matrix of each cluster region, and selecting a region with an average temperature lower than 260K and an energy value greater than 0.8 as a candidate region.

[0009] Preferably, the selection of the region with the lowest average temperature from the candidate regions as the gas reference region, and the calculation of the average multispectral vector of the gas reference region as the gas reference spectrum comprise: traversing all candidate regions, calculating the average temperature of all pixel points in each candidate region; sorting all candidate regions in ascending order according to the average temperature; selecting the first region after sorting as the gas reference region; extracting the response values of all pixel points in the gas reference region in N spectral channels, calculating the average response value of each spectral channel to form an average multispectral vector of N dimensions, and taking the average multispectral vector as the gas reference spectrum.

[0010] Preferably, the step of calculating the within-group variance of each spectral channel within the gas reference region and determining the weight vector based on the within-group variance includes: calculating the variance of the spectral response values ​​of all pixels in the i-th spectral channel within the gas reference region. Calculate the variance of the spectral response values ​​of all pixels in N spectral channels within the gas reference region to obtain the variance set. Weight of the i-th spectral channel Calculated using the following formula: ; The weights of all spectral channels constitute an N-dimensional weight vector. .

[0011] Preferably, the angle threshold is calculated in the following way: The aggregate variance value is obtained by summing the variances within all spectral channel groups. ; The angle threshold is calculated using the following formula. ;in, For angle threshold, This is the proportionality coefficient. This is the aggregate variance value. Based on the offset.

[0012] Preferably, for any pixel to be detected in the frame image, calculating the weighted spectral angle between the multispectral vector of the pixel to be detected and the gas reference spectrum based on the weight vector includes: The multispectral vector of the pixel to be detected is , Gas reference spectrum The weight vector is ; Weighted spectral angle Calculated using the following formula .

[0013] Preferably, the step of determining a pixel to be detected as a gas leak point if it simultaneously meets the following conditions: the weighted spectral angle is less than an angle threshold, and the temperature of the pixel to be detected is lower than a preset temperature threshold, includes: The preset temperature threshold is 265K; Obtain the weighted spectral angle of the pixel to be detected and temperature ; Determine whether the conditions are met simultaneously. and ; If both conditions are met, the state of the pixel to be detected is marked as 1; otherwise, it is marked as 0.

[0014] Preferably, generating a visual image of the gas leak area based on all determined gas leak points includes: creating a zero-mask image with the same size as the original single-frame image; traversing each pixel in the original single-frame image and checking the determination state of the obtained pixel; if the state of a pixel is 1, then setting the pixel value at the corresponding position in the zero-mask image to a preset color to generate a mask image with the leak area; and overlaying the generated mask image with the leak area with the original infrared image with transparency to obtain a visual image.

[0015] Preferably, acquiring the infrared thermal imaging multispectral sequence image of the ring main unit includes: acquiring the infrared thermal imaging multispectral sequence image of the ring main unit using a cooled multispectral infrared thermal imager.

[0016] Secondly, the sulfur hexafluoride ring main unit gas leakage detection system based on thermal imaging of the present invention includes a memory and a processor. The memory stores computer instructions, and when the processor executes the computer instructions, it implements the above-mentioned sulfur hexafluoride ring main unit gas leakage detection method based on thermal imaging.

[0017] The beneficial effects of this invention are as follows: This invention can automatically identify and extract the gas spectrum at the scene from infrared multispectral images as a real-time reference, avoiding the environmental mismatch problem caused by the reliance on a fixed spectral library in traditional methods, thus improving the applicability of detection. By combining spectral features and temperature features for dual joint discrimination, the unique spectral absorption characteristics of the gas are combined with the low-temperature physical phenomena during leakage, which can eliminate interference from low-temperature objects or background reflections in the environment that only meet a single feature, reducing the false alarm rate. During spectral matching, different recognition weights are assigned to different spectral channels based on the stability of the extracted gas reference spectrum itself, and a judgment threshold related to the quality of the reference spectrum is constructed. This allows the detection process to focus on the most stable characteristic bands of the gas and adjust the discrimination scale according to the clarity of the on-site signal, thereby achieving more reliable identification of sulfur hexafluoride gas leaks in complex backgrounds and improving the anti-interference capability of detection. Attached Figure Description

[0018] Figure 1 This is a schematic flowchart of a gas leakage detection method for a sulfur hexafluoride ring main unit based on thermal imaging, provided in an embodiment of the present invention. Detailed Implementation

[0019] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0020] like Figure 1As shown, an embodiment of the sulfur hexafluoride ring main unit gas leakage detection method based on thermal imaging provided by the present invention includes the following steps: S1. Acquire infrared thermal imaging multispectral sequence images of the ring main unit; identify multiple candidate regions with low temperature and uniform texture in each frame of the multispectral sequence images through mean-shift clustering; select the region with the lowest average temperature from the candidate regions as the gas reference region, and calculate the average multispectral vector of the gas reference region as the gas reference spectrum.

[0021] Specifically, a cooled multispectral infrared thermal imager is used to acquire a sequence of infrared thermal imaging multispectral images of the ring main unit. The cooled multispectral infrared thermal imager has a built-in filter wheel or tunable filter, enabling it to acquire images of multiple narrowband spectral channels in the mid-wave or long-wave infrared bands, including at least the strong absorption band of sulfur hexafluoride gas near 10.55 micrometers. The cooled multispectral infrared thermal imager is aimed at the area of ​​the ring main unit to be inspected, resulting in a time-series multispectral image data cube. Each frame of the image contains radiation intensity information in different spectral channels at that moment. A single frame of multispectral image is converted into a single-channel temperature image, and a multidimensional feature vector containing spatial coordinate information, temperature information, and texture information is constructed for each pixel in the image.

[0022] Preferably, the identification of multiple candidate regions with low temperature and uniform texture in each frame of a multispectral sequence image through mean-shift clustering includes: S11: Initialize cluster centers by using all pixels in the frame image as the initial sample point set; S12: Set spatial domain bandwidth The color gamut bandwidth is 15. It is 20; S13: Calculate the mean shift vector for each sample point within the bandwidth range, and move the sample point to the end of the mean shift vector. S14: Repeat step S13 until the moving distance of all sample points is less than the preset convergence threshold, and the clustering is completed. S15: Calculate the average temperature and energy value of the gray-level co-occurrence matrix for each cluster region, and select regions with an average temperature below 260K and an energy value greater than 0.8 as candidate regions.

[0023] For example, all 327,680 pixels in a 640×512 pixel infrared image are used as the initial sample point set. Each pixel consists of a two-dimensional spatial coordinate and a temperature value, forming a three-dimensional feature vector. If the spatial search radius is 15 pixels and the temperature search range is 20 units, then for any pixel, such as the one located at coordinates 100, 100, the mean drift vector of all sample points within its 15-pixel neighborhood with a temperature difference less than 20 is calculated, and the sample point is moved to the endpoint of the mean drift vector. This moving process is repeated for all pixels until the moving distance of each pixel in the iteration is less than a preset convergence threshold. At this point, the pixels converge to several density centers, forming clustered regions.

[0024] After clustering, assuming the image is divided into five distinct regions, the average temperature and energy value of the gray-level co-occurrence matrix (GLCM) of each region are analyzed. For example, region 1 has an average temperature of 255K and an energy value of 0.9; region 2 has an average temperature of 290K and an energy value of 0.7; and region 3 has an average temperature of 258K and an energy value of 0.85. Based on the selection criteria, regions 1 and 3 are candidate regions. High energy values ​​indicate that the pixel gray-level distribution in that region is uniform, the texture is simple, and it matches the background characteristics.

[0025] Preferably, the step of selecting the region with the lowest average temperature from the candidate regions as the gas reference region and calculating the average multispectral vector of the gas reference region as the gas reference spectrum includes: Traverse all candidate regions and calculate the average temperature of all pixels within each candidate region; Sort all candidate regions in ascending order based on their average temperature. The first region after sorting is selected as the gas reference region; Extract the response values ​​of all pixels in the gas reference region on N spectral channels, calculate the average response value of each spectral channel, and construct an N-dimensional average multispectral vector. The average multispectral vector is used as the gas reference spectrum.

[0026] For example, as described above, two candidate regions, Region 1 and Region 3, were obtained. The average temperature of Region 1 was calculated to be 255K, and the average temperature of Region 3 was 258K. The two regions were sorted in ascending order of average temperature, resulting in a list in the order of Region 1, Region 3. The first region in the list, which has the lowest average temperature, Region 1, was selected as the gas reference region. Region 1 best represents the low-temperature background without gas interference in the current scene. After selecting Region 1, assuming the multispectral imaging device has 8 channels, i.e., N equals 8, all pixels in Region 1 were traversed, and the response value of each pixel on the 8 spectral channels was extracted. The average response value was calculated independently for each channel. For example, the average response value of all pixels in Region 1 for Channel 1 is 150.3, the average response value for Channel 2 is 180.5, and so on up to Channel 8. The 8 average values ​​constitute an 8-dimensional average multispectral vector, which serves as the gas reference spectrum.

[0027] S2, calculate the within-group variance of each spectral channel within the gas reference region, and determine the weight vector and angle threshold based on the within-group variance; for any pixel to be detected in the frame image, calculate the weighted spectral angle between the multispectral vector of the pixel to be detected and the gas reference spectrum based on the weight vector.

[0028] In one embodiment, calculating the within-group variance of each spectral channel within the gas reference region and determining the weight vector based on the within-group variance includes: Calculate the variance of the spectral response values ​​of all pixels in the i-th spectral channel within the gas reference region. ; Calculate the variance of the spectral response values ​​of all pixels in N spectral channels within the gas reference region to obtain the variance set. ; Weight of the i-th spectral channel Calculated using the following formula: ; The weights of all spectral channels constitute an N-dimensional weight vector. .

[0029] For example, as described earlier, in a gas reference region using 8 spectral channels, firstly, for channel 1, the variance of the response values ​​of all pixels in that gas reference region on channel 1 is calculated. A smaller variance indicates a more stable signal and lower noise in the channel. Assume the calculated variance for channel 1 is 2.0. This process is repeated for the remaining 7 channels, resulting in a variance set containing 8 variance values, for example (2.0, 5.0, 1.5, 8.0, 4.0, 6.0, 2.5, 9.0). After obtaining the variance set, a weight is calculated for each spectral channel. Spectral channels with more stable signals receive a larger weight, while those with higher noise levels receive a lower weight.

[0030] In one embodiment, the angle threshold is calculated in the following manner: The aggregate variance value is obtained by summing the variances within all spectral channel groups. ; The angle threshold is calculated using the following formula. ;in, For angle threshold, This is the proportionality coefficient. This is the aggregate variance value. Based on the offset.

[0031] In one embodiment, calculating the weighted spectral angle between the multispectral vector of the pixel to be detected and the gas reference spectrum based on the weight vector for any pixel to be detected in the frame image includes: The multispectral vector of the pixel to be detected is , Gas reference spectrum The weight vector is ; Weighted spectral angle Calculated using the following formula .

[0032] For example, taking the pixel with coordinates (200, 350) as an example, the 8-channel multispectral vector of this pixel is extracted. Assuming multispectral vectors The gas reference spectrum is (145.1, 182.3, ...). The weight vector is (150.3, 180.5, ...). The values ​​are (0.2, 0.08, ...). The weighted spectral angle is obtained using the weighted spectral angle calculation formula. By using weighted spectral angles, the similarity between the spectrum of the pixel under test and the gas reference spectrum can be determined, and the influence of the stable channel is enhanced by the weight vector, while the interference of the noise channel is suppressed.

[0033] S3, determine the pixel to be detected as a gas leak point if it meets the following conditions: the weighted spectral angle is less than the angle threshold and the temperature of the pixel to be detected is lower than the preset temperature threshold; generate a visual image of the gas leak area based on all the determined gas leak points.

[0034] In one embodiment, determining a pixel to be detected as a gas leak point if it simultaneously meets the following conditions: the weighted spectral angle is less than an angle threshold, and the temperature of the pixel to be detected is lower than a preset temperature threshold, includes: The preset temperature threshold is 265K; Obtain the weighted spectral angle of the pixel to be detected and temperature ; Determine whether the conditions are met simultaneously. and ; If both conditions are met, the state of the pixel to be detected is marked as 1; otherwise, it is marked as 0.

[0035] A pixel's status is marked as 1, indicating that the pixel is a gas leak point; a pixel's status is marked as 0, indicating that the pixel is not a leak point.

[0036] In one embodiment, generating a visual image of the gas leak area based on all identified gas leak points includes: Create a zero-mask image with the same dimensions as the original single-frame image; Iterate through each pixel in the original single-frame image and check the determination status of the obtained pixel. If the state of a pixel is 1, the pixel value at the corresponding position in the all-zero mask image is set to the preset color to generate a mask image with a leakage area; The generated mask image with the leaked area is overlaid with the original infrared image to obtain a visual image.

[0037] For example, after determining the status of all pixels in an original single-frame image, a zero-value mask image with the same size as the original single-frame image is created. Initially, all pixels in the zero-value mask image have a value of 0. Each pixel in the original single-frame image is then iterated through, and its status is checked. For instance, if a pixel at coordinates (200, 350) is found and its status is marked as 1, a preset color, such as a semi-transparent red, is drawn at position (200, 350) in the zero-value mask image.

[0038] After inspecting all pixels in the original single-frame image, all locations identified as gas leak points on the all-zero mask image are filled with semi-transparent red, forming color blocks indicating the shape and extent of the gas cloud, while other locations remain transparent, thus generating a mask image with the leak area. This generated mask image with the leak area is then overlaid with the original infrared image using transparency adjustments to obtain a visualized image.

[0039] The implementation principle of the sulfur hexafluoride (SF6) ring main unit gas leak detection method based on thermal imaging in this invention is as follows: This invention can automatically identify and extract the real-time spectrum of the gas directly from the infrared multispectral image at the site as a reference, thereby avoiding the environmental mismatch problem caused by the reliance on a fixed spectral library in traditional methods, making the detection more universal. Furthermore, this invention combines the spectral characteristics and temperature characteristics of the gas for "dual joint" discrimination. By simultaneously verifying the gas's unique spectral absorption effect and the low-temperature phenomenon generated during leakage, it can effectively eliminate interference items in the environment that only satisfy a single characteristic, such as low-temperature objects or background reflection, significantly reducing the false alarm rate. Moreover, this invention can dynamically assign recognition weights to different spectral channels based on the stability of the extracted reference spectrum itself, and construct a judgment threshold related to spectral quality. This allows the detection process to automatically focus on the most stable and clearest characteristic bands of the gas, and adjust the "tightness" of the discrimination according to the signal quality, thus enabling this invention to reliably identify SF6 gas leaks even in complex backgrounds and with weak signals.

[0040] An embodiment of the sulfur hexafluoride ring main unit gas leak detection system based on thermal imaging provided by the present invention includes a memory and a processor. The memory stores computer instructions, and when the processor executes the computer instructions, it implements the sulfur hexafluoride ring main unit gas leak detection method based on thermal imaging in the above embodiment.

[0041] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for detecting gas leakage in a sulfur hexafluoride ring main unit based on thermal imaging, characterized in that, Includes the following steps: S1. Acquire infrared thermal imaging multispectral sequence images of the ring main unit; identify multiple candidate regions with low temperature and uniform texture in each frame of the multispectral sequence images through mean-shift clustering; select the region with the lowest average temperature from the candidate regions as the gas reference region, and calculate the average multispectral vector of the gas reference region as the gas reference spectrum. S2, calculate the within-group variance of each spectral channel within the gas reference region, and determine the weight vector and angle threshold based on the within-group variance; for any pixel to be detected in the frame image, calculate the weighted spectral angle between the multispectral vector of the pixel to be detected and the gas reference spectrum based on the weight vector; S3, determine the pixel to be detected as a gas leak point if it meets the following conditions: the weighted spectral angle is less than the angle threshold and the temperature of the pixel to be detected is lower than the preset temperature threshold; generate a visual image of the gas leak area based on all the determined gas leak points.

2. The method for detecting gas leakage in a sulfur hexafluoride ring main unit based on thermal imaging according to claim 1, characterized in that, The method of identifying multiple candidate regions with low temperature and uniform texture in each frame of a multispectral sequence image through mean-shift clustering includes: S11: Initialize cluster centers by using all pixels in the frame image as the initial sample point set; S12: Set spatial domain bandwidth The color gamut bandwidth is 15. It is 20; S13: Calculate the mean shift vector for each sample point within the bandwidth range, and move the sample point to the end of the mean shift vector. S14: Repeat step S13 until the moving distance of all sample points is less than the preset convergence threshold, and the clustering is completed. S15: Calculate the average temperature and energy value of the gray-level co-occurrence matrix for each cluster region, and select regions with an average temperature below 260K and an energy value greater than 0.8 as candidate regions.

3. The method for detecting gas leakage in a sulfur hexafluoride ring main unit based on thermal imaging according to claim 1, characterized in that, The step of selecting the region with the lowest average temperature from the candidate regions as the gas reference region and calculating the average multispectral vector of the gas reference region as the gas reference spectrum includes: Traverse all candidate regions and calculate the average temperature of all pixels within each candidate region; Sort all candidate regions in ascending order based on their average temperature. The first region after sorting is selected as the gas reference region; Extract the response values ​​of all pixels in the gas reference region on N spectral channels, calculate the average response value of each spectral channel, and construct an N-dimensional average multispectral vector. The average multispectral vector is used as the gas reference spectrum.

4. The method for detecting gas leakage in a sulfur hexafluoride ring main unit based on thermal imaging according to claim 1, characterized in that, The calculation of the within-group variance of each spectral channel within the gas reference region, and the determination of the weight vector based on the within-group variance, includes: Calculate the variance of the spectral response values ​​of all pixels in the i-th spectral channel within the gas reference region. ; Calculate the variance of the spectral response values ​​of all pixels in N spectral channels within the gas reference region to obtain the variance set. ; Weight of the i-th spectral channel Calculated using the following formula: ； The weights of all spectral channels constitute an N-dimensional weight vector. .

5. The method for detecting gas leakage in a sulfur hexafluoride ring main unit based on thermal imaging according to claim 1, characterized in that, The angle threshold is calculated in the following way: The aggregate variance value is obtained by summing the variances within all spectral channel groups. ; The angle threshold is calculated using the following formula. ;in, For angle threshold, This is the proportionality coefficient. This is the aggregate variance value. Based on the offset.

6. The method for detecting gas leakage in a sulfur hexafluoride ring main unit based on thermal imaging according to claim 1, characterized in that, For any pixel to be detected in the frame image, the weighted spectral angle between the multispectral vector of the pixel to be detected and the gas reference spectrum is calculated based on the weight vector, including: The multispectral vector of the pixel to be detected is , Gas reference spectrum The weight vector is ; Weighted spectral angle Calculated using the following formula 。 7. The method for detecting gas leakage in a sulfur hexafluoride ring main unit based on thermal imaging according to claim 1, characterized in that, The step of identifying a pixel as a gas leak point if it simultaneously meets the following conditions: the weighted spectral angle is less than an angle threshold, and the temperature of the pixel is lower than a preset temperature threshold, includes: The preset temperature threshold is 265K; Obtain the weighted spectral angle of the pixel to be detected and temperature ; Determine whether the conditions are met simultaneously. and ; If both conditions are met, the state of the pixel to be detected is marked as 1; otherwise, it is marked as 0.

8. The method for detecting gas leakage in a sulfur hexafluoride ring main unit based on thermal imaging according to claim 1, characterized in that, The step of generating a visual image of the gas leak area based on all identified gas leak points includes: Create a zero-mask image with the same dimensions as the original single-frame image; Iterate through each pixel in the original single-frame image and check the determination status of the obtained pixel. If the state of a pixel is 1, the pixel value at the corresponding position in the all-zero mask image is set to the preset color to generate a mask image with a leakage area; The generated mask image with the leaked area is overlaid with the original infrared image to obtain a visual image.

9. The method for detecting gas leakage in a sulfur hexafluoride ring main unit based on thermal imaging according to claim 1, characterized in that, The acquisition of infrared thermal imaging multispectral sequence images of the ring main unit includes: acquiring infrared thermal imaging multispectral sequence images of the ring main unit using a cooled multispectral infrared thermal imager.

10. A sulfur hexafluoride ring main unit gas leak detection system based on thermal imaging, characterized in that, It includes a memory and a processor. The memory stores computer instructions. When the processor executes the computer instructions, it implements the gas leakage detection method for sulfur hexafluoride ring main unit based on thermal imaging as described in any one of claims 1-9.

Images