A method and system for imaging and locating SF6 gas leakage in a confined space

Abstract

A method and system for imaging and locating SF6 gas leakage in a confined space, belonging to the technical field of SF6 leakage detection in confined spaces, solves the industry problem of difficulty in quickly locating SF6 gas leakage points in confined spaces; the technical solution of the present invention first displays the leaked invisible SF6 gas imaging, then records the spatial coordinates of the SF6 gas telemetry field of view coaxial alignment structure and SF6 cloud alarm image data, and then uses a binocular three-dimensional positioning and reconstruction method to calculate the three-dimensional spatial coordinates of the SF6 gas leakage point, and combines with a patrol robot to realize large-scale, high-precision patrol monitoring and leakage imaging and positioning functions, filling the technical gap in this field and providing precise coordinates for the next step of recycling and processing.

Description

Technical Field

[0001] This invention belongs to the field of SF6 leakage detection technology in confined spaces, and relates to an imaging and localization method and system for SF6 gas leakage in confined spaces. Background Technology

[0002] Enclosed spaces in the power industry mainly include transmission line corridors, underground substations, and indoor substations. These enclosed spaces have a small footprint and strong environmental adaptability, making them widely used in densely populated urban areas, areas with complex geographical environments, and other similar locations. According to statistics, the State Grid Corporation currently operates over 8,000 indoor substations and over 50 underground substations. Integrated utility tunnel projects, exemplified by the Sutong GIL (Gas Insulated Line) project, are also emerging and developing across various regions.

[0003] Electrical equipment insulated with SF6 (Sulphur Hexafluoride) gas is characterized by its compact structure, small size, electromagnetically friendly environment, and high safety. It is widely used in confined electrical spaces; the Sutong GIL alone used over 800 tons of SF6 gas. As an insulating and arc-quenching medium in equipment, a large leak of SF6 gas will inevitably lead to a decline in the equipment's insulation performance and cause operational failures. Furthermore, SF6 gas has a greenhouse effect more than 23,900 times that of CO2, can exist in the air for over 3,200 years, and is one of the six gases banned from emission under the Kyoto Protocol of the United Nations Framework Convention on Climate Change. SF6 is five times denser than air and tends to accumulate in low-lying areas. Pure SF6 gas is asphyxiating, and SF6 gas after operation may contain highly toxic decomposition products. Therefore, a large-scale SF6 gas leak in a confined space can easily cause mass casualties and have serious social consequences.

[0004] Currently, both domestic and international methods for monitoring SF6 gas leaks in confined spaces employ a distributed sensor approach. Sensors are placed at different locations within the confined space, and the sensors only respond when the leaking gas diffuses to the sensor sites and accumulates to a certain concentration. Therefore, this method heavily relies on the gas diffusion rate and concentration, resulting in problems such as data lag, low sensitivity, and inaccurate location. Furthermore, the sensors are prone to aging and failure, leading to a short lifespan. In 2016, the State Grid Anhui Electric Power Company, during a performance survey of indoor SF6 leak alarm devices, found that the pass rate of leak monitoring sensors installed in indoor substations was less than 30%. Therefore, there is an urgent need to research new technologies for SF6 gas leak monitoring and develop a novel intelligent monitoring system for SF6 gas leaks in confined spaces to achieve early warning and precise location of SF6 leaks.

[0005] In existing technologies, the literature "Using Laser Imaging Technology to Locate and Detect Gas Leaks in SF6 Equipment" (Tian Yong et al., Northeast Electric Power Technology, 2005), published in 2015, describes a laser imaging leak detector that emits infrared laser light to the equipment under test and scans it. If there is no leak, the reflected infrared light energy remains unchanged; if a leak exists, the reflected infrared light energy decreases, and the video image captured by the instrument will show a smoke-like shadow. The higher the SF6 gas concentration, the more laser energy is absorbed, and the more obvious the smoke-like shadow becomes. Based on these factors, the presence of a gas leak, the leak point, the direction of gas movement, and the leak speed can be determined. This detection method has a short detection time and can achieve visualized live detection, but it is easily affected by external airflow and humidity, resulting in poor detection stability.

[0006] Currently, the most advanced on-site detection method is long-distance, non-contact, spectral technology represented by Fourier transform infrared spectroscopy telemetry. This not only improves accuracy and detection sensitivity through spectral analysis, but also facilitates large-scale autonomous inspection of various scenarios through passive telemetry, greatly meeting the actual needs of SF6 leak detection in power scenarios. However, how to quickly locate SF6 gas leak points remains a challenge for the industry. Summary of the Invention

[0007] The purpose of this invention is to design an imaging and positioning method and system for SF6 gas leaks in confined spaces, so as to solve the industry problem of rapid positioning of SF6 gas leak points in confined spaces.

[0008] The present invention solves the above-mentioned technical problems through the following technical solutions:

[0009] A method for imaging and locating SF6 gas leaks in a confined space includes the following steps:

[0010] S1. SF6 Gas Leakage Distribution Imaging and Display Processing: An SF6 gas telemetry field-of-view coaxial night vision camera is used to acquire scene images of a confined space, which are then used as the background image for the SF6 gas leakage distribution. Pixels identified as SF6 gas are synthesized into a pseudo-color image of SF6 gas concentration distribution according to quantitative concentration ranges. Based on the field-of-view coaxial pixel registration mapping relationship, the pseudo-color image of SF6 gas concentration distribution is fused with the background image to form an SF6 gas leakage distribution map.

[0011] S2. Detecting SF6 gas leaks and alarming with first and second spatial coordinates: Mount the SF6 gas telemetry device on the inspection robot and move along the set route. When the SF6 gas telemetry device detects an SF6 gas leak and alarms, check the displayed position of the SF6 gas cloud on the image. If the SF6 gas cloud is located at the set position in the image, record the current spatial coordinates of the SF6 gas telemetry device; otherwise, continue moving until the coordinate recording condition is met. The SF6 gas telemetry device continues moving until it detects another SF6 gas leak at the same location and the SF6 gas cloud is located at a position symmetrical to the set position in the image. Record the current spatial coordinates of the SF6 gas telemetry device; otherwise, continue moving until the coordinate recording condition is met.

[0012] S3. Location Reconstruction: Using the spatial coordinates of the SF6 gas telemetry instrument and the alarm image data of the SF6 gas cloud, the three-dimensional spatial coordinates of the SF6 gas leak point are calculated and saved using the binocular three-dimensional location reconstruction method.

[0013] First, the invisible leaked SF6 gas is imaged and displayed. Then, the spatial coordinates of the SF6 gas telemetry instrument and the alarm image data of the SF6 cloud are recorded. Next, the three-dimensional spatial coordinates of the SF6 gas leak point are calculated using a binocular three-dimensional positioning and reconstruction method. Combined with an inspection robot, this method can achieve large-scale, high-precision inspection and monitoring and leak imaging and positioning functions, filling the technological gap in this field and providing accurate coordinates for the next step of recycling and processing.

[0014] As a further improvement to the technical solution of the present invention, the SF6 gas telemetry field of view coaxial registration structure includes: a coaxial registration bracket (1), a night vision camera (2), an infrared focal plane detector (3), an infrared focal plane detector base (4), set screws (5), and set screws (6); the coaxial registration bracket (1) and the infrared focal plane detector base (4) are both fixed on the same horizontal plane, the night vision camera (2) is installed inside the coaxial registration bracket (1), a hole with the same shape as the lens of the night vision camera (2) is opened on the front panel of the coaxial registration bracket (1), the lens of the night vision camera (2) is installed in the hole, two set screws (5) are set at the upper left and lower right corners of the hole, and two set screws (6) are set at the lower left and upper right corners of the hole, the two set screws (5) and the two set screws (6) are distributed at the four vertices of the square, and the center of the square is the same as the night vision camera. The centers of the lenses of the night vision camera (2) coincide, and two set screws (5) and two set screws (6) are used to adjust the field of view of the night vision camera (2); the infrared focal plane detector (3) is fixed on the infrared focal plane detector base (4); the lens of the night vision camera (2) and the lens of the infrared focal plane detector (3) are set at the same height and coaxially, and the distance between the lens of the night vision camera (2) and the lens of the infrared focal plane detector (3) is d; if the field of view of the night vision camera (2) is tilted to the left, the upper left corner set screw (5) and the lower left corner set screw (6) are unscrewed; if the field of view of the night vision camera (2) is tilted to the right, the lower right corner set screw (5) and the upper right corner set screw (6) are unscrewed; if the field of view of the night vision camera (2) is tilted upward, the upper left corner set screw (5) and the upper right corner set screw (6) are unscrewed; if the field of view of the night vision camera (2) is tilted downward, the lower right corner set screw (5) and the lower left corner set screw (6) are unscrewed.

[0015] As a further improvement to the technical solution of the present invention, the method for identifying pixels as SF6 gas is as follows:

[0016] S11. Use an infrared focal plane detector (3) to obtain the multi-element correction score map of the leaked SF6 gas;

[0017] S12. Use the surrounding pixel average filling method to correct the bad pixels in the SF6 gas multivariate correction score map caused by damage to some pixels or abnormal signal of the infrared focal plane detector (3).

[0018] S13. Perform smoothing filtering on the SF6 gas multivariate correction score map, and calculate the average center intensity of the background region and the SF6 gas region to facilitate image segmentation.

[0019] S14. Segment the SF6 gas multivariate correction score map to obtain the background region and the SF6 gas region.

[0020] S15. Perform signal-to-noise ratio enhancement on the SF6 gas multivariate correction score map, converting the SF6 gas multivariate correction score map into a signal-to-noise ratio image.

[0021] S16. Calculate the gradient of the signal-to-noise ratio image and remove edge pixels simultaneously;

[0022] S17. Perform local threshold segmentation on the already segmented background region and SF6 gas region;

[0023] S18. Based on the local threshold segmentation results, the set of pixels with values ​​greater than the threshold is taken as the SF6 gas region, thus completing the identification of SF6 gas.

[0024] As a further improvement to the technical solution of the present invention, the registration method of the SF6 gas telemetry field-of-view coaxial registration structure includes the following steps:

[0025] S111, Image scaling factor calibration: Using the checkerboard pattern target board as the background, the night vision camera (2) and the infrared focal plane detector (3) respectively acquire the checkerboard board image, select a certain checkerboard grid, calculate the total number of pixels of the two images in the grid, and record them as NP. visible and NP infrared Calculate the infrared image scaling factor Q = NP visible / NP infrared After obtaining the scaling factor, the entire infrared image is scaled by Q.

[0026] S112, Coaxial Adjustment and Registration: Select the sun as the target at infinity, move the entire coaxial registration structure, first adjust the central axis of the field of view of the infrared focal plane detector (3) to align with the sun, then adjust the set screw (5) and set screw (6) to align the central axis of the field of view of the night vision camera (2) with the sun. When the central axes of the fields of view of the night vision camera (2) and the infrared focal plane detector (3) are both aligned with the target at infinity, it is considered that the fields of view of the night vision camera (2) and the infrared focal plane detector (3) are coaxial. Since there is a distance d between the optical axes of the night vision camera (2) and the infrared focal plane detector (3), the pixels are translated to make the target pattern on the images of the two coincide, thus completing the coaxial registration of the fields of view.

[0027] As a further improvement to the technical solution of the present invention, the method of adjusting the set screw (5) and the set screw (6) to align the central axis of the field of view of the night vision camera (2) with the sun is as follows: if the field of view of the night vision camera (2) is tilted to the left, the set screw (5) at the upper left corner and the set screw (6) at the lower left corner are unscrewed; if the field of view of the night vision camera (2) is tilted to the right, the set screw (5) at the lower right corner and the set screw (6) at the upper right corner are unscrewed; if the field of view of the night vision camera (2) is tilted upward, the set screw (5) at the upper left corner and the set screw (6) at the upper right corner are unscrewed; if the field of view of the night vision camera (2) is tilted downward, the set screw (5) at the lower right corner and the set screw (6) at the lower left corner are unscrewed.

[0028] An imaging and localization system for SF6 gas leaks in a confined space includes: an imaging display and processing module, a spatial coordinate detection module, and a localization and reconstruction module;

[0029] The imaging display processing module uses an SF6 gas telemetry field-of-view coaxial night vision camera to acquire scene images of a confined space and uses them as a background image of SF6 gas leakage distribution; it synthesizes pixels identified as SF6 gas into a pseudo-color image of SF6 gas concentration distribution according to quantitative concentration ranges; and it fuses the pseudo-color image of SF6 gas concentration distribution with the background image according to the field-of-view coaxial pixel registration mapping relationship to form an SF6 gas leakage distribution map.

[0030] The aforementioned spatial coordinate detection module mounts an SF6 gas telemetry device onto an inspection robot, which then moves along a pre-defined route. When the SF6 gas telemetry device detects an SF6 gas leak and triggers an alarm, it checks the position of the SF6 gas cloud on the image. If the SF6 gas cloud is located at the pre-defined position in the image, the spatial coordinates of the SF6 gas telemetry device are recorded; otherwise, the robot continues moving until the coordinate recording condition is met. The SF6 gas telemetry device continues moving until it detects another SF6 gas leak at the same location, and the SF6 gas cloud is located symmetrically to the pre-defined position in the image. If so, the spatial coordinates of the SF6 gas telemetry device are recorded; otherwise, the robot continues moving until the coordinate recording condition is met.

[0031] The positioning and reconstruction module utilizes the spatial coordinates of the SF6 gas telemetry instrument and the alarm image data of the SF6 gas cloud, which are recorded sequentially, to calculate and save the three-dimensional spatial coordinates of the SF6 gas leak point using a binocular three-dimensional positioning and reconstruction method.

[0032] As a further improvement to the technical solution of the present invention, the SF6 gas telemetry field of view coaxial registration structure includes: a coaxial registration bracket (1), a night vision camera (2), an infrared focal plane detector (3), an infrared focal plane detector base (4), set screws (5), and set screws (6); the coaxial registration bracket (1) and the infrared focal plane detector base (4) are both fixed on the same horizontal plane, the night vision camera (2) is installed inside the coaxial registration bracket (1), a hole with the same shape as the lens of the night vision camera (2) is opened on the front panel of the coaxial registration bracket (1), the lens of the night vision camera (2) is installed in the hole, two set screws (5) are set at the upper left and lower right corners of the hole, and two set screws (6) are set at the lower left and upper right corners of the hole, the two set screws (5) and the two set screws (6) are distributed at the four vertices of the square, and the center of the square is the same as the night vision camera. The centers of the lenses of the night vision camera (2) coincide, and two set screws (5) and two set screws (6) are used to adjust the field of view of the night vision camera (2); the infrared focal plane detector (3) is fixed on the infrared focal plane detector base (4); the lens of the night vision camera (2) and the lens of the infrared focal plane detector (3) are set at the same height and coaxially, and the distance between the lens of the night vision camera (2) and the lens of the infrared focal plane detector (3) is d; if the field of view of the night vision camera (2) is tilted to the left, the upper left corner set screw (5) and the lower left corner set screw (6) are unscrewed; if the field of view of the night vision camera (2) is tilted to the right, the lower right corner set screw (5) and the upper right corner set screw (6) are unscrewed; if the field of view of the night vision camera (2) is tilted upward, the upper left corner set screw (5) and the upper right corner set screw (6) are unscrewed; if the field of view of the night vision camera (2) is tilted downward, the lower right corner set screw (5) and the lower left corner set screw (6) are unscrewed.

[0033] As a further improvement to the technical solution of the present invention, the method for identifying pixels as SF6 gas is as follows:

[0034] S11. Use an infrared focal plane detector (3) to obtain the multi-element correction score map of the leaked SF6 gas;

[0035] S12. Use the surrounding pixel average filling method to correct the bad pixels in the SF6 gas multivariate correction score map caused by damage to some pixels or abnormal signal of the infrared focal plane detector (3).

[0036] S13. Perform smoothing filtering on the SF6 gas multivariate correction score map, and calculate the average center intensity of the background region and the SF6 gas region to facilitate image segmentation.

[0037] S14. Segment the SF6 gas multivariate correction score map to obtain the background region and the SF6 gas region.

[0038] S15. Perform signal-to-noise ratio enhancement on the SF6 gas multivariate correction score map, converting the SF6 gas multivariate correction score map into a signal-to-noise ratio image.

[0039] S16. Calculate the gradient of the signal-to-noise ratio image and remove edge pixels simultaneously;

[0040] S17. Perform local threshold segmentation on the already segmented background region and SF6 gas region;

[0041] S18. Based on the local threshold segmentation results, the set of pixels with values ​​greater than the threshold is taken as the SF6 gas region, thus completing the identification of SF6 gas.

[0042] As a further improvement to the technical solution of the present invention, the registration method of the SF6 gas telemetry field-of-view coaxial registration structure includes the following steps:

[0043] S111, Image scaling factor calibration: Using the checkerboard pattern target board as the background, the night vision camera (2) and the infrared focal plane detector (3) respectively acquire the checkerboard board image, select a certain checkerboard grid, calculate the total number of pixels of the two images in the grid, and record them as NP. visible and NP infrared Calculate the infrared image scaling factor Q = NP visible / NP infrared After obtaining the scaling factor, the entire infrared image is scaled by Q.

[0044] S112, Coaxial Adjustment and Registration: Select the sun as the target at infinity, move the entire coaxial registration structure, first adjust the central axis of the field of view of the infrared focal plane detector (3) to align with the sun, then adjust the set screw (5) and set screw (6) to align the central axis of the field of view of the night vision camera (2) with the sun. When the central axes of the fields of view of the night vision camera (2) and the infrared focal plane detector (3) are both aligned with the target at infinity, it is considered that the fields of view of the night vision camera (2) and the infrared focal plane detector (3) are coaxial. Since there is a distance d between the optical axes of the night vision camera (2) and the infrared focal plane detector (3), the pixels are translated to make the target pattern on the images of the two coincide, thus completing the coaxial registration of the fields of view.

[0045] As a further improvement to the technical solution of the present invention, the method of adjusting the set screw (5) and the set screw (6) to align the central axis of the field of view of the night vision camera (2) with the sun is as follows: if the field of view of the night vision camera (2) is tilted to the left, the set screw (5) at the upper left corner and the set screw (6) at the lower left corner are unscrewed; if the field of view of the night vision camera (2) is tilted to the right, the set screw (5) at the lower right corner and the set screw (6) at the upper right corner are unscrewed; if the field of view of the night vision camera (2) is tilted upward, the set screw (5) at the upper left corner and the set screw (6) at the upper right corner are unscrewed; if the field of view of the night vision camera (2) is tilted downward, the set screw (5) at the lower right corner and the set screw (6) at the lower left corner are unscrewed.

[0046] The advantages of this invention are:

[0047] (1) The technical solution of the present invention first processes the image display of the leaked invisible SF6 gas, then records the spatial coordinates of the SF6 gas telemetry instrument and the alarm image data of the SF6 cloud, and then uses the binocular three-dimensional positioning and reconstruction method to calculate the three-dimensional spatial coordinates of the SF6 gas leak point. Combined with the inspection robot, it can realize the large-scale, high-precision inspection monitoring and leak imaging positioning function, fill the technical gap in this field, and provide accurate coordinates for the next step of recycling and processing.

[0048] (2) The SF6 gas telemetry field of view coaxial registration structure designed in this invention realizes the optical axis registration between the night vision camera and the infrared focal plane detector base through the field of view adjustment structure, providing hardware support for remote imaging and positioning of SF6 gas leakage in confined spaces and large-area scanning scene image stitching. The coaxial registration structure of this invention is small in size, light in weight and low in cost, which greatly improves the cost performance of SF6 gas leakage detection. Attached Figure Description

[0049] Figure 1 This is a flowchart of an imaging and localization method for SF6 gas leakage in a confined space according to an embodiment of the present invention;

[0050] Figure 2 This is an imaging display of SF6 gas distribution according to an embodiment of the present invention;

[0051] Figure 3 This is a front view of the coaxial registration structure for SF6 gas telemetry field of view according to an embodiment of the present invention;

[0052] Figure 4 This is a front view of the coaxial registration structure for SF6 gas telemetry field of view according to an embodiment of the present invention;

[0053] Figure 5 This is a schematic diagram of image scaling factor calibration for the SF6 gas telemetry field-of-view coaxial registration method according to an embodiment of the present invention;

[0054] Figure 6 This is a schematic diagram of coaxial adjustment and registration of the SF6 gas telemetry field of view coaxial registration method according to an embodiment of the present invention;

[0055] Figure 7 This is a flowchart of the method for identifying pixels as SF6 gas according to an embodiment of the present invention;

[0056] Figure 8 This is a diagram showing the recognition result of the SF6 gas multivariate correction score map according to an embodiment of the present invention;

[0057] Figure 9 This is a schematic diagram of the SF6 gas leak binocular three-dimensional localization and reconstruction method according to an embodiment of the present invention. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] The technical solution of the present invention will be further described below with reference to the accompanying drawings and specific embodiments:

[0060] Example 1

[0061] like Figure 1 As shown, a method for imaging and locating SF6 gas leaks in a confined space includes the following steps:

[0062] 1. Imaging and display processing of SF6 gas leak distribution

[0063] 1) Use a coaxial night vision camera with SF6 gas telemetry field of view to acquire scene images of the confined space and use them as background images of SF6 gas leakage distribution;

[0064] 2) Pixels identified as SF6 gas are synthesized into a pseudo-color image of SF6 gas concentration distribution according to quantitative concentration ranges;

[0065] 3) Based on the coaxial pixel registration mapping relationship of the field of view, the pseudo-color image of SF6 gas concentration distribution is fused with the background image to form an SF6 gas leakage distribution map, such as... Figure 2 As shown.

[0066] like Figure 3-4As shown, the SF6 gas telemetry field-of-view coaxial registration structure includes: a coaxial registration bracket 1, a night vision camera 2, an infrared focal plane detector 3, an infrared focal plane detector base 4, set screws 5, and set screws 6. The coaxial registration bracket 1 and the infrared focal plane detector base 4 are both fixed on the same horizontal plane. The night vision camera 2 is installed inside the coaxial registration bracket 1. A hole identical to the lens of the night vision camera 2 is opened on the front panel of the coaxial registration bracket 1, and the lens of the night vision camera 2 is installed in the hole. Two screws are provided at the upper left and lower right corners of the hole. A set screw 5 is provided, and two set screws 6 are provided at the lower left and upper right corners of the hole. The two set screws 5 and two set screws 6 are distributed at the four vertices of the square. The center of the square coincides with the center of the lens of the night vision camera 2. The two set screws 5 and two set screws 6 are used to adjust the field of view of the night vision camera 2. The infrared focal plane detector 3 is fixed on the infrared focal plane detector base 4. The lens of the night vision camera 2 and the lens of the infrared focal plane detector 3 are set at the same height and coaxially. The distance between the lens of the night vision camera 2 and the lens of the infrared focal plane detector 3 is d.

[0067] The method for adjusting the field of view of night vision camera 2 using set screw 5 and set screw 6 is as follows:

[0068] 1) If the field of view of night vision camera 2 is tilted to the left, unscrew the upper left corner set screw 5 and the lower left corner set screw 6;

[0069] 2) If the field of view of night vision camera 2 is tilted to the right, unscrew the lower right corner set screw 5 and the upper right corner set screw 6;

[0070] 3) If the field of view of night vision camera 2 is tilted upwards, unscrew the upper left corner set screw 5 and the upper right corner set screw 6;

[0071] 4) If the field of view of night vision camera 2 is tilted downwards, unscrew the set screw 5 in the lower right corner and the set screw 6 in the lower left corner.

[0072] The registration method for the SF6 gas telemetry field-of-view coaxial registration structure is as follows:

[0073] a. Image scaling factor calibration

[0074] Because the single-pixel field of view of night vision camera 2 and infrared focal plane detector 3 are different, the image scaling factors of the two need to be calibrated before field of view registration to ensure that the same target has a consistent size in the image. Figure 5 As shown, with a checkerboard patterned target board as the background, night vision camera 2 and infrared focal plane detector 3 respectively acquire images of the checkerboard board. A specific checkerboard square is selected, and the total number of pixels in the square from both images is calculated, denoted as NP. visible and NP infrared Calculate the infrared image scaling factor Q = NP visible / NP infraredAfter obtaining the scaling factor, the entire infrared image is scaled by Q.

[0075] b. Coaxial adjustment and registration

[0076] like Figure 6 As shown, the sun is selected as the target at infinity. The entire coaxial registration structure is moved around. First, the central axis of the infrared focal plane detector 3 is aligned with the sun. Then, the set screw 5 and set screw 6 are adjusted to align the central axis of the night vision camera 2 with the sun. When both the central axes of the night vision camera 2 and the infrared focal plane detector 3 are aligned with the target at infinity, the fields of view of the night vision camera 2 and the infrared focal plane detector 3 are considered to be coaxial. Since there is a distance d between the optical axes of the night vision camera 2 and the infrared focal plane detector 3, pixels need to be translated to make the target image on both images coincide. This completes the coaxial registration of the fields of view.

[0077] like Figure 7 As shown, the method for identifying pixels as SF6 gas is as follows:

[0078] 1) Obtaining the SF6 score map: The infrared focal plane detector directly obtains the SF6 multivariate correction results in the SF6 gas remote sensing field of view coaxial registration structure device, which is called the SF6 multivariate correction score map;

[0079] 2) Bad Pixel Correction: For bad pixels in the SF6 multivariate correction score map caused by partial pixel damage or signal abnormality in the infrared focal plane detector, the bad pixels are corrected by the average filling method of surrounding pixels.

[0080] 3) Signal averaging: Perform general smoothing filtering on the score image, and calculate the average center intensity of the background region and the SF6 gas region to facilitate the next step of image segmentation;

[0081] 4) Image segmentation: The score image is segmented using a general image segmentation method to obtain the background region and the SF6 gas region;

[0082] 5) Signal-to-noise ratio enhancement: The MRCE algorithm is used to enhance the signal-to-noise ratio of the score image, transforming it into a signal-to-noise ratio image;

[0083] The MRCE algorithm works as follows: Low global contrast between the target and background, and the transformation of the target's local neighborhood into a target region with high contrast between the target and background, are addressed by MRCE. MRCE utilizes a series of different sliding modules as filters on each pixel of the image to exclude the local neighborhood of the measured pixel, enhance the globally blurred region of interest, and convert the "scoring image" into a "signal-to-noise ratio image" by using local intensity statistics to express the intensity of the scoring image. The value of each pixel in the signal-to-noise ratio image represents the highest local signal-to-noise ratio value calculated across all modules.

[0084] 6) Image gradient calculation: A general image gradient calculation method is used to calculate the gradient of the signal-to-noise ratio image, while removing edge pixels;

[0085] 7) Local threshold setting: Perform local threshold segmentation on the already segmented background area and SF6 gas area;

[0086] 8) SF6 identification: Based on the local threshold segmentation results, the set of pixels larger than the threshold is regarded as the SF6 gas region, and the SF6 gas identification is completed.

[0087] like Figure 8 As shown, the method for identifying pixels as SF6 gas processes the obtained SF6 multivariate correction score map, which can effectively suppress noisy pixels in the SF6 multivariate correction score map. While suppressing noise, it improves the accuracy of effective signal recognition, reduces false alarm and false alarm rates, and improves the SF6 gas leak detection effect.

[0088] 2. SF6 gas leak detection alarm: first spatial coordinates and second spatial coordinates.

[0089] like Figure 9 As shown in the figure, L(X,Y,Z) are the spatial coordinates of the gas leak point, C1 is the image plane of the SF6 gas telemetry instrument before movement, C2 is the image plane of the SF6 gas telemetry instrument after movement, O1 is the infrared telemetry coordinate system before movement, O2 is the infrared telemetry coordinate system after movement, a1(u,v) is the projection of the gas leak point on the image plane of the SF6 gas telemetry instrument before movement, and a2(u,v) is the projection of the gas leak point on the image plane of the SF6 gas telemetry instrument after movement.

[0090] The SF6 gas telemetry device is mounted on the inspection robot and moves along a predetermined route. When the SF6 gas telemetry device detects an SF6 gas leak and alarms, it checks the position of the SF6 cloud on the image. If the SF6 cloud is located in the middle 1 / 3 of the right side of the image, the current spatial coordinates of the telemetry device are recorded. Otherwise, it continues to move forward until the coordinate recording conditions are met.

[0091] The SF6 gas telemetry instrument continues to advance several meters until it detects another SF6 gas leak at the same location and the SF6 cloud is located in the middle 1 / 3 of the left side of the image. If so, it records the current spatial coordinates of the telemetry instrument. Otherwise, it continues to advance until the coordinate recording conditions are met.

[0092] Using the spatial coordinates of the SF6 gas telemetry instrument and the alarm image data of the SF6 cloud, the three-dimensional spatial coordinates of the SF6 gas leak point are calculated using a binocular three-dimensional positioning and reconstruction method. Traditional binocular three-dimensional positioning and reconstruction methods can be found in "Research on Depth Perception and Positioning Technology Based on Binocular Vision" (Beijing University of Posts and Telecommunications, February 2020, Deng Xinyi). Traditional binocular three-dimensional positioning and reconstruction methods locate the three-dimensional spatial coordinates of visible objects. However, the leaked SF6 gas in this invention is invisible, requiring imaging and display processing of the leaked SF6 gas to convert it into a visible target before using the binocular three-dimensional positioning and reconstruction method for three-dimensional spatial coordinate positioning.

[0093] 3. Location Reconstruction

[0094] Using the spatial coordinates of the SF6 gas telemetry instrument and the alarm image data of the SF6 gas cloud, the three-dimensional spatial coordinates of the SF6 gas leak point were calculated and saved by using a binocular three-dimensional positioning and reconstruction method.

[0095] The spatial coordinates of the gas leak point L were calculated using the coordinate relationship before and after the SF6 gas telemetry instrument moved, as well as the camera parameters.

[0096] 1) SF6 gas telemetry camera calibration: After the SF6 gas telemetry is moved, the positional change will inevitably cause the image planes C1 and C2 to be non-parallel. Calibration is required to make the two image planes on the same plane and each row in the image have the same direction and coordinates.

[0097] 2) 3D reprojection: Based on the parallax of the gas leak point, the geometric position of the SF6 gas telemetry instrument, and the camera focal length parameters, the distance from the SF6 gas telemetry instrument to the gas leak point is calculated using triangulation, and the spatial coordinates of the gas leak point are calculated.

[0098] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for imaging and locating SF6 gas leakage in a confined space, characterized in that: The following steps are involved: S1. SF6 gas leakage distribution imaging display processing: Use the SF6 gas telemetry field of view coaxial night vision camera to obtain the scene image of the confined space, and use it as the background image of the SF6 gas leakage distribution; synthesize the pixels identified as SF6 gas into a SF6 gas concentration distribution pseudo-color image according to the quantitative concentration interval; merge the SF6 gas concentration distribution pseudo-color image and the background image into an SF6 gas leakage distribution map according to the field of view coaxial pixel registration mapping relationship; S2. Detect the first spatial coordinate and the second spatial coordinate of the SF6 gas leakage alarm: carry the SF6 gas telemeter on the inspection robot and move forward along the set route. When the SF6 gas telemeter finds SF6 gas leakage and alarms, check the display position of the SF6 gas cloud on the image. If the SF6 gas cloud is located at the set position in the image, record the spatial coordinates of the current SF6 gas telemeter; otherwise, continue to move forward until the coordinate recording condition is met; the SF6 gas telemeter continues to move forward until SF6 gas leakage is found at the same position again and the SF6 gas cloud is located at a position symmetrical to the set position in the image, then record the spatial coordinates of the current SF6 gas telemeter; otherwise, continue to move forward until the coordinate recording condition is met; S3. Positioning and reconstruction: Using the spatial coordinates of the SF6 gas telemeter recorded successively and the SF6 gas cloud alarm image data, a binocular three-dimensional positioning and reconstruction method is adopted to calculate the three-dimensional spatial coordinates of the SF6 gas leakage point and save the spatial coordinates of the gas leakage point.

2. A method for imaging and locating SF6 gas leakage in a confined space according to claim 1, characterized in that: The SF6 gas telemetry field of view coaxial alignment structure comprises: a coaxial alignment bracket (1), a night vision camera (2), an infrared focal plane detector (3), an infrared focal plane detector base (4), a set screw (5), and a top screw (6); the coaxial alignment bracket (1) and the infrared focal plane detector base (4) are fixed on the same horizontal plane, the night vision camera (2) is installed inside the coaxial alignment bracket (1), a hole identical to the lens of the night vision camera (2) is opened on the front panel of the coaxial alignment bracket (1), the lens of the night vision camera (2) is installed in the hole, two set screws (5) are arranged at the upper left corner and the lower right corner of the hole, two top screws (6) are arranged at the lower left corner and the upper right corner of the hole, the two set screws (5) and the two top screws (6) are distributed at the four vertices of a square, and the center of the square is aligned with the lens of the night vision camera (2). The centers coincide, and two set screws (5) and two top screws (6) are used to adjust the field of view of the night vision camera (2); the infrared focal plane detector (3) is fixed on the infrared focal plane detector base (4); the lens of the night vision camera (2) and the lens of the infrared focal plane detector (3) are coaxially arranged at the same height, and the distance between the lens of the night vision camera (2) and the lens of the infrared focal plane detector (3) is d; if the field of view of the night vision camera (2) deviates to the left, the upper left set screw (5) and the lower left top screw (6) are screwed out; if the field of view of the night vision camera (2) deviates to the right, the lower right set screw (5) and the upper right top screw (6) are screwed out; if the field of view of the night vision camera (2) deviates upward, the upper left set screw (5) and the upper right top screw (6) are screwed out; if the field of view of the night vision camera (2) deviates downward, the lower right set screw (5) and the lower left top screw (6) are screwed out.

3. A method for imaging and locating SF6 gas leakage in a confined space according to claim 2, characterized in that: The method for identifying the pixel as SF6 gas is as follows: S11, using an infrared focal plane detector (3) to obtain a multivariate correction score diagram of the leaked SF6 gas; S12, using the surrounding pixel average filling method to correct the bad pixels in the SF6 gas multivariate correction score map caused by partial pixel damage or signal abnormality of the infrared focal plane detector (3); S13, performing smoothing filtering on the SF6 gas multivariate correction score map, and calculating the regional average central intensity for the background area and the SF6 gas area, so as to facilitate image segmentation; S14, segmenting the SF6 gas multivariate correction score map to obtain a background area and a SF6 gas area; S15, performing signal-to-noise ratio enhancement on the SF6 gas multivariate correction score map, and converting the SF6 gas multivariate correction score map into a signal-to-noise ratio image, S16, performing gradient calculation on the signal-to-noise ratio image and removing edge pixels; S17, performing local threshold segmentation on the segmented background area and SF6 gas area; S18. According to the local threshold segmentation result, a set of pixel points greater than the threshold is taken as the SF6 gas area to complete the identification of SF6 gas.

4. A method for imaging and locating SF6 gas leakage in a confined space according to claim 2, characterized in that: The alignment method of the SF6 gas telemetry field of view coaxial alignment structure comprises the following steps: S111, image scaling factor calibration: With the chessboard pattern target board as the background, the night vision camera (2) and the infrared focal plane detector (3) respectively obtain the chessboard image, select a chessboard grid, and calculate the total number of pixels of the two images in the grid, which are recorded as NP visible and NP infrared , calculate the infrared image scaling factor Q = NP visible / NP infrared , after obtaining the scaling factor, scale the entire infrared image by Q; S112, coaxial adjustment and alignment: select the sun as the infinitely distant alignment target, move the entire coaxial alignment structure, first adjust the field of view center axis of the infrared focal plane detector (3) to align with the sun, then adjust the set screw (5) and the top screw (6) to align the field of view center axis of the night vision camera (2) with the sun. When the fields of view center axes of the night vision camera (2) and the infrared focal plane detector (3) are both aligned with the infinitely distant target, it is considered that the fields of view of the night vision camera (2) and the infrared focal plane detector (3) are coaxial. Since there is a distance d between the optical axes of the night vision camera (2) and the infrared focal plane detector (3), the pixels are translated so that the target graphics on the two images overlap, and the field of view coaxial alignment is completed.

5. A method for imaging and locating SF6 gas leakage in a confined space according to claim 4, characterized in that: The method for adjusting the set screw (5) and the top screw (6) so that the central axis of the field of view of the night vision camera (2) is aligned with the sun is specifically as follows: if the field of view of the night vision camera (2) deviates to the left, the set screw (5) at the upper left corner and the top screw (6) at the lower left corner are screwed out; if the field of view of the night vision camera (2) deviates to the right, the set screw (5) at the lower right corner and the top screw (6) at the upper right corner are screwed out; if the field of view of the night vision camera (2) deviates to the upper part, the set screw (5) at the upper left corner and the top screw (6) at the upper right corner are screwed out; if the field of view of the night vision camera (2) deviates to the lower part, the set screw (5) at the lower right corner and the top screw (6) at the lower left corner are screwed out.

6. A confined space SF6 gas leakage imaging and positioning system, characterized in that: include: Imaging display processing module, spatial coordinate detection module, positioning reconstruction module; The imaging display processing module uses a SF6 gas telemetry field of view coaxial night vision camera to obtain a scene image of a confined space, and uses it as a background image of SF6 gas leakage distribution; the pixels identified as SF6 gas are synthesized into a pseudo-color image of SF6 gas concentration distribution according to a quantitative concentration interval; the pseudo-color image of SF6 gas concentration distribution is merged with the background image into a SF6 gas leakage distribution map according to a field of view coaxial pixel registration mapping relationship; The spatial coordinate detection module carries the SF6 gas telemeter on the inspection robot and moves along the set route. When the SF6 gas telemeter finds SF6 gas leakage and alarms, it checks the display position of the SF6 gas cloud on the image. If the SF6 gas cloud is located at the set position in the image, the spatial coordinates of the current SF6 gas telemeter are recorded; otherwise, it continues to move forward until the coordinate recording condition is met; the SF6 gas telemeter continues to move forward until SF6 gas leakage is found at the same position again and the SF6 gas cloud is located at a position symmetrical to the set position in the image, then the spatial coordinates of the current SF6 gas telemeter are recorded; otherwise, it continues to move forward until the coordinate recording condition is met; The positioning and reconstruction module uses the spatial coordinates of the SF6 gas telemeter and the SF6 gas cloud alarm image data recorded successively, adopts a binocular three-dimensional positioning and reconstruction method, calculates the three-dimensional spatial coordinates of the SF6 gas leakage point and saves the spatial coordinates of the gas leakage point.

7. The confined space SF6 gas leakage imaging and positioning system according to claim 6, characterized in that: The SF6 gas telemetry field of view coaxial alignment structure comprises: a coaxial alignment bracket (1), a night vision camera (2), an infrared focal plane detector (3), an infrared focal plane detector base (4), a set screw (5), and a top screw (6); the coaxial alignment bracket (1) and the infrared focal plane detector base (4) are fixed on the same horizontal plane, the night vision camera (2) is installed inside the coaxial alignment bracket (1), a hole identical to the lens of the night vision camera (2) is opened on the front panel of the coaxial alignment bracket (1), the lens of the night vision camera (2) is installed in the hole, two set screws (5) are arranged at the upper left corner and the lower right corner of the hole, two top screws (6) are arranged at the lower left corner and the upper right corner of the hole, the two set screws (5) and the two top screws (6) are distributed at the four vertices of a square, and the center of the square is aligned with the lens of the night vision camera (2). The centers coincide, and two set screws (5) and two top screws (6) are used to adjust the field of view of the night vision camera (2); the infrared focal plane detector (3) is fixed on the infrared focal plane detector base (4); the lens of the night vision camera (2) and the lens of the infrared focal plane detector (3) are coaxially arranged at the same height, and the distance between the lens of the night vision camera (2) and the lens of the infrared focal plane detector (3) is d; if the field of view of the night vision camera (2) deviates to the left, the upper left set screw (5) and the lower left top screw (6) are screwed out; if the field of view of the night vision camera (2) deviates to the right, the lower right set screw (5) and the upper right top screw (6) are screwed out; if the field of view of the night vision camera (2) deviates upward, the upper left set screw (5) and the upper right top screw (6) are screwed out; if the field of view of the night vision camera (2) deviates downward, the lower right set screw (5) and the lower left top screw (6) are screwed out.

8. The confined space SF6 gas leakage imaging and positioning system according to claim 7, characterized in that: The method for identifying the pixel as SF6 gas is as follows: S11, using an infrared focal plane detector (3) to obtain a multivariate correction score diagram of the leaked SF6 gas; S12, using the surrounding pixel average filling method to correct the bad pixels in the SF6 gas multivariate correction score map caused by partial pixel damage or signal abnormality of the infrared focal plane detector (3); S13, performing smoothing filtering on the SF6 gas multivariate correction score map, and calculating the regional average central intensity for the background area and the SF6 gas area, so as to facilitate image segmentation; S14, segmenting the SF6 gas multivariate correction score map to obtain a background area and a SF6 gas area; S15, performing signal-to-noise ratio enhancement on the SF6 gas multivariate correction score map, and converting the SF6 gas multivariate correction score map into a signal-to-noise ratio image, S16, performing gradient calculation on the signal-to-noise ratio image and removing edge pixels; S17, performing local threshold segmentation on the segmented background area and SF6 gas area; S18. According to the local threshold segmentation result, a set of pixel points greater than the threshold is taken as the SF6 gas area to complete the identification of SF6 gas.

9. The confined space SF6 gas leakage imaging and positioning system according to claim 7, characterized in that: The alignment method of the SF6 gas telemetry field of view coaxial alignment structure comprises the following steps: S111, image scaling factor calibration: With the chessboard pattern target board as the background, the night vision camera (2) and the infrared focal plane detector (3) respectively obtain the chessboard image, select a chessboard grid, and calculate the total number of pixels of the two images in the grid, which are recorded as NP visible and NP infrared , calculate the infrared image scaling factor Q = NP visible / NP infrared , after obtaining the scaling factor, scale the entire infrared image by Q; S112, coaxial adjustment and alignment: select the sun as the infinitely distant alignment target, move the entire coaxial alignment structure, first adjust the field of view center axis of the infrared focal plane detector (3) to align with the sun, then adjust the set screw (5) and the top screw (6) to align the field of view center axis of the night vision camera (2) with the sun. When the fields of view center axes of the night vision camera (2) and the infrared focal plane detector (3) are both aligned with the infinitely distant target, it is considered that the fields of view of the night vision camera (2) and the infrared focal plane detector (3) are coaxial. Since there is a distance d between the optical axes of the night vision camera (2) and the infrared focal plane detector (3), the pixels are translated so that the target graphics on the two images overlap, and the field of view coaxial alignment is completed.

10. The confined space SF6 gas leakage imaging and positioning system according to claim 9, characterized in that: The method for adjusting the set screw (5) and the top screw (6) so that the central axis of the field of view of the night vision camera (2) is aligned with the sun is specifically as follows: if the field of view of the night vision camera (2) deviates to the left, the set screw (5) at the upper left corner and the top screw (6) at the lower left corner are screwed out; if the field of view of the night vision camera (2) deviates to the right, the set screw (5) at the lower right corner and the top screw (6) at the upper right corner are screwed out; if the field of view of the night vision camera (2) deviates to the upper part, the set screw (5) at the upper left corner and the top screw (6) at the upper right corner are screwed out; if the field of view of the night vision camera (2) deviates to the lower part, the set screw (5) at the lower right corner and the top screw (6) at the lower left corner are screwed out.

Images