Method and system for early warning of stress concentration hot spots of hydrogen storage reactor based on machine vision

Abstract

The present application relates to the field of image processing, in particular to a hydrogen storage reactor stress concentration hotspot early warning method and system based on machine vision, comprising: according to the pressure of the place where the hydrogen storage reactor is located when each image is shot, the real-time temperature of the area where the camera is located, the real-time temperature of each pixel point in each speckle pattern in the image corresponding to the surface area of the hydrogen storage reactor, the central wavelength of the filter added in front of the camera lens, and the depth information of each pixel point, the stress concentration index of each pixel point in each speckle pattern in each image is obtained, and the inflation process of the hydrogen storage reactor is warned. The present application aims to solve the problem that the existing method has relatively low warning accuracy when detecting the stress of the hydrogen storage reactor during the inflation process, because there is a relatively large difference between the temperature around the hydrogen storage reactor and the temperature around the camera.

Description

Technical Field

[0001] This invention relates to the field of image processing, and more specifically to a method and system for early warning of stress concentration hotspots in hydrogen storage reactors based on machine vision. Background Technology

[0002] As the energy structure undergoes a deep decarbonization transition, hydrogen energy, as a secondary energy carrier with extremely high gravimetric energy density and zero carbon emissions, is gradually being applied in multiple core areas such as transportation, industrial manufacturing, and grid energy storage. During operation, high-pressure hydrogen gas is forced into the hydrogen storage reactor within a very short time. Due to factors such as the adiabatic compression effect of the gas and the kinetic energy dissipation caused by gas velocity conversion, a dramatic heat accumulation occurs inside the reactor. This rapid temperature rise not only causes significant thermal expansion of the inner liner material but also induces thermal mismatch stress at the interface between the inner liner and the outer carbon fiber composite material. Achieving non-contact, real-time early deformation monitoring and warning of stress concentration hotspots on the surface of the hydrogen storage reactor before macroscopic cracks form and lead to pressure leakage has become a critical technical challenge in the safe operation and maintenance of hydrogen energy equipment.

[0003] As a non-contact, full-field deformation measurement method, the core principle of this algorithm is to acquire random speckle patterns on the surface of the object being measured using a binocular stereo industrial camera, and then use computer vision algorithms to track the spatial displacement of each small subset of pixels before and after the object is loaded. Finally, through complex coordinate transformations and numerical differentiation, the displacement and strain of the object's surface are analytically determined. However, existing methods... When the algorithm is directly ported to the hydrogen charging and discharging scenario of a high-pressure hydrogen storage reactor, the video monitoring system will be significantly affected by the underlying physical environment in the current scenario, making it difficult to... When the algorithm provides early warnings for hydrogen storage reactors, the accuracy of the warning results is relatively low. Summary of the Invention

[0004] This invention provides a machine vision-based method and system for early warning of stress concentration hotspots in hydrogen storage reactors, to solve the problems of existing... When the algorithm is directly applied to the hydrogen charging and discharging scenario of a high-pressure hydrogen storage reactor, the accuracy of the early warning results is relatively low.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A speckle pattern was sprayed onto the surface of the hydrogen storage reactor. When the hydrogen storage reactor began to be filled with hydrogen, the surface of the hydrogen storage reactor was photographed at fixed intervals to obtain multiple images.

[0007] Based on the pressure at the location of the hydrogen storage reactor when each image is captured, the real-time temperature of the area where the camera is located, the real-time temperature of the surface area of ​​the hydrogen storage reactor corresponding to each pixel in each speckle pattern in the image, the center wavelength of the filter installed in front of the camera lens, and the depth information of each pixel in the image, the distortion compensation weight of each pixel in each speckle pattern in each image is obtained.

[0008] Based on the distortion compensation weights of each pixel within each speckle pattern in each image, its position within the image, and using... The algorithm initializes the subset side length of each pixel and obtains the adaptive subset side length of each pixel within each speckle pattern in each image.

[0009] Based on the adaptive subset side length of each pixel in each speckle pattern in each image, the stretching ratio of each pixel in each speckle pattern in each image is obtained; based on the stretching ratio of each pixel in each speckle pattern in each image, the adaptive subset side length and the stretching ratio of the surrounding pixels, the stress concentration index of each pixel in each speckle pattern in each image is obtained.

[0010] The gasification process of the hydrogen storage reactor is monitored based on the internal pressure and shape of the reactor when each image is acquired, as well as the stress concentration index of each pixel within each speckle pattern in each image.

[0011] Furthermore, the specific calculation steps for spraying a speckle pattern onto the surface of the hydrogen storage reactor and taking multiple images of the reactor surface at fixed intervals when hydrogen is started to be introduced into the reactor are as follows:

[0012] A high-contrast speckle pattern is sprayed onto the surface of the hydrogen storage reactor. A binocular stereo industrial camera is set up with a set acquisition interval. When the hydrogen storage reactor starts to be filled with hydrogen, the surface of the hydrogen storage reactor is photographed at the set acquisition interval to obtain multiple images. When photographing the hydrogen storage reactor, the optical axis of the binocular stereo industrial camera is aligned with the central axis of the hydrogen storage reactor tank.

[0013] Furthermore, the specific calculation steps for obtaining the distortion compensation weight of each pixel in each speckle pattern in each image based on the pressure of the hydrogen storage reactor location when each image is captured, the real-time temperature of the area where the camera is located, the real-time temperature of the surface area of ​​the hydrogen storage reactor corresponding to each pixel in each speckle pattern in the image, the center wavelength of the filter installed in front of the camera lens, and the depth information of each pixel in the image are as follows:

[0014] An infrared thermal imager array is deployed to obtain the real-time temperature of the surface area of ​​the hydrogen storage reactor corresponding to each pixel in each speckle pattern in the image when each image is captured, and to obtain the real-time temperature of the area where the binocular stereo industrial camera used for capturing the image is located when each image is captured.

[0015] The pressure at the location of the hydrogen storage reactor was obtained by using atmospheric pressure sensors placed at the site of the hydrogen storage reactor when each image was taken.

[0016] Get the The first image The first speckle pattern within the first The specific formula for calculating the distortion compensation weight of each pixel is as follows:

[0017]

[0018] In the formula, Indicates the first The first image The first speckle pattern within the first Distortion compensation weights for each pixel Gladstone constant representing air. Indicates the shooting of the first The image shows the pressure at the location of the hydrogen storage reactor. This represents the specific gas constant of dry air. Indicates the shooting of the first The image shows the real-time temperature of the area where the binocular stereo industrial camera used for taking the picture. Indicates the shooting of the first The first image The first image The first speckle pattern within the first Each pixel corresponds to the real-time temperature of the surface area of ​​the hydrogen storage reactor. Represents the absolute value function. This indicates the center wavelength of the narrowband filter installed in front of the lens of a stereo industrial camera used for shooting. Indicates the first The first image The first speckle pattern within the first Depth information of each pixel.

[0019] Furthermore, the distortion compensation weights for each pixel within each speckle pattern in each image, its position within the image, and the use of... The algorithm initializes the subset side length of pixels and calculates the adaptive subset side length of each pixel within each speckle pattern in each image using the following steps:

[0020] The first Using the centroid of the hydrogen storage reactor in the image as the origin, a two-dimensional coordinate system is constructed with a straight line parallel to the central axis of the hydrogen storage reactor tank as the vertical axis and a straight line perpendicular to the vertical axis as the horizontal axis.

[0021] The first The first image The first speckle pattern within the first The value of the nth pixel on the horizontal axis is denoted as the nth pixel. The first image The first speckle pattern within the first The distance between each pixel and the central axis of the hydrogen storage reactor;

[0022] Get and use The initial values ​​of the subset side lengths of pixels in the algorithm, and the radius of the cylindrical section of the hydrogen storage reactor when it is not filled with gas;

[0023] According to usage The algorithm initializes the side length of the subset of pixels and the first... The first image The first speckle pattern within the first The distortion compensation weights of each pixel, the distance from the central axis of the hydrogen storage reactor, the radius of the cylindrical section of the hydrogen storage reactor when it is not filled with gas, and the physical spatial resolution parameters of the binocular stereo industrial camera used for shooting after calibration on the front surface of the hydrogen storage reactor are used to obtain the first pixel. The first image The first speckle pattern within the first Adaptive subset side length for each pixel.

[0024] Furthermore, the statement based on use The algorithm initializes the side length of the subset of pixels and the first... The first image The first speckle pattern within the first The distortion compensation weights of each pixel, the distance from the central axis of the hydrogen storage reactor, the radius of the cylindrical section of the hydrogen storage reactor when it is not filled with gas, and the physical spatial resolution parameters of the binocular stereo industrial camera used for shooting after calibration on the front surface of the hydrogen storage reactor are used to obtain the first pixel. The first image The first speckle pattern within the first The specific steps for calculating the adaptive subset side length of each pixel are as follows:

[0025]

[0026] In the formula, Indicates the first The first image The first speckle pattern within the first Adaptive subset side length of each pixel Indicates use The algorithm initializes the side length of a subset of pixels. Indicates the first The first image The first speckle pattern within the first Distortion compensation weights for each pixel This indicates the radius of the cylindrical section of the hydrogen storage reactor when it is not filled with gas. This represents the physical spatial resolution parameters of the binocular stereo industrial camera used for filming, after calibration on the front surface of the hydrogen storage reactor. Indicates the first The first image The first speckle pattern within the first The distance between each pixel and the central axis of the hydrogen storage reactor.

[0027] Furthermore, the specific calculation steps for obtaining the stretching ratio of each pixel in each speckle pattern in each image based on the adaptive subset side length of each pixel in each speckle pattern in each image are as follows:

[0028] Using the DIC algorithm, according to the first... The first image The first speckle pattern within the first The adaptive subset side length of the nth pixel is obtained to get the nth pixel. The first image The first speckle pattern within the first The displacement value of the nth pixel, according to the nth The displacement values ​​of all pixels in the image are used to obtain the first... The first image The first speckle pattern within the first Stretching rate per pixel.

[0029] Furthermore, the specific calculation steps for obtaining the stress concentration index of each pixel in each speckle pattern in each image based on the stretching ratio of each pixel in each speckle pattern in each image, the adaptive subset side length, and the stretching ratio of surrounding pixels are as follows:

[0030] Will be located in the The first image The first speckle pattern within the first The pixel within the eight-neighborhood of the i-th pixel, and located within the speckle pattern, is denoted as the i-th pixel. The first image The first speckle pattern within the first The neighboring pixels of a pixel;

[0031] Get the The first image The first speckle pattern within the first The specific formula for calculating the stress concentration index of each pixel is as follows:

[0032]

[0033] In the formula, Indicates the first The first image The first speckle pattern within the first Stress concentration index per pixel Indicates the first The mean stretching ratio of pixels within all speckle patterns in the image. Indicates the first The first image The first speckle pattern within the first Strength per pixel Indicates the first The first image The first speckle pattern within the first The average stretching ratio of all neighboring pixels of a given pixel. Indicates the first The first image The first speckle pattern within the first Adaptive subset side length of each pixel This represents the physical spatial resolution parameters of the binocular stereo industrial camera used for filming, after calibration on the front surface of the hydrogen storage reactor. It is an exponential function with the natural constant as its base. It represents 1 meter.

[0034] Furthermore, the specific calculation steps for providing early warning of the hydrogen storage reactor's charging process based on the internal pressure of the hydrogen storage reactor, the shape of the hydrogen storage reactor, and the stress concentration index of each pixel within each speckle pattern in each image are as follows:

[0035] Obtain the internal pressure of the hydrogen storage reactor before it is filled with hydrogen, and record it as the initial pressure.

[0036] The internal pressure of the hydrogen storage reactor increased from the initial pressure when hydrogen was introduced into the reactor. The mean stress concentration index of all pixels within all speckle patterns in all images acquired at that time is recorded as the stress concentration index of the internal pressure of the hydrogen storage reactor rising from the initial pressure to the value obtained when hydrogen is introduced into the reactor. The average stress concentration index of the entire tank of the hydrogen storage reactor. ;

[0037] Based on the pressure inside the hydrogen storage reactor when each image was acquired, the shape of the hydrogen storage reactor, and the increase in internal pressure from the initial pressure to the pressure when hydrogen was introduced into the hydrogen storage reactor. The stress concentration index of the entire tank of the hydrogen storage reactor was obtained. The warning index of the image;

[0038] If the first In the image, the stress concentration index of pixels located within the speckle pattern is greater than that of the first pixel. The warning index of the first image determines the warning level at the time of collection. When the image is captured, abnormal stress deformation caused by local microscopic defects and structural damage on the surface of the hydrogen storage reactor is detected. At this time, the early warning program is triggered and a warning is reported to the terminal.

[0039] Furthermore, the method involves considering the pressure inside the hydrogen storage reactor when acquiring each image, the shape of the hydrogen storage reactor, and the increase in internal pressure from the initial pressure to the pressure when hydrogen is introduced into the hydrogen storage reactor. The stress concentration index of the entire tank of the hydrogen storage reactor was obtained. The specific steps for calculating the warning index of Zhang image are as follows:

[0040]

[0041] In the formula, Indicates the first The warning index of the image, This indicates that when hydrogen is introduced into the hydrogen storage reactor, the internal pressure rises from the initial pressure to... The average stress concentration index of the entire tank of the hydrogen storage reactor. This indicates the design axis length of the tank body of the hydrogen storage reactor. This indicates the radius of the cylindrical section of the hydrogen storage reactor when it is not filled with gas. This indicates the rated permissible pressure for the hydrogen storage reactor to operate at its limit. Indicates obtaining the first The pressure inside the hydrogen storage reactor when the image is taken. Let be a logarithmic function with the natural constant as its base. It is a natural constant.

[0042] A second aspect of the present invention provides a machine vision-based early warning system for stress concentration hotspots in a hydrogen storage reactor. This system includes a data acquisition module, a distortion compensation weight calculation module, an adaptive subset side length module, a stress concentration index calculation module, and an early warning module, wherein:

[0043] The data acquisition module is used to spray a speckle pattern on the surface of the hydrogen storage reactor. When the hydrogen storage reactor starts to be filled with hydrogen, the surface of the hydrogen storage reactor is photographed at fixed intervals to obtain multiple images.

[0044] The distortion compensation weight calculation module is used to obtain the distortion compensation weight of each pixel in each speckle pattern in each image based on the pressure of the hydrogen storage reactor when each image is captured, the real-time temperature of the area where the camera is located, the real-time temperature of the surface area of ​​the hydrogen storage reactor corresponding to each pixel in each speckle pattern in the image, the center wavelength of the filter installed in front of the camera lens, and the depth information of each pixel in the image.

[0045] The adaptive subset side length module is used to determine the distortion compensation weights of each pixel within each speckle pattern in each image, its position within the image, and the use of... The algorithm initializes the subset side length of each pixel and obtains the adaptive subset side length of each pixel within each speckle pattern in each image.

[0046] The stress concentration index calculation module is used to obtain the stretching ratio of each pixel in each speckle pattern in each image based on the adaptive subset side length of each pixel in each speckle pattern in each image; and to obtain the stress concentration index of each pixel in each speckle pattern in each image based on the stretching ratio of each pixel in each speckle pattern in each image, the adaptive subset side length and the stretching ratio of the surrounding pixels.

[0047] The early warning module is used to provide early warnings about the gasification process of the hydrogen storage reactor based on the pressure inside the hydrogen storage reactor, the shape of the hydrogen storage reactor, and the stress concentration index of each pixel in each speckle pattern in each image when each image is acquired.

[0048] The beneficial effects of the technical solution of the present invention are:

[0049] 1. Overcoming the coupling interference of thermal flow disturbance and curvature distortion: This invention introduces an infrared thermal imager and an environmental temperature and pressure sensor, and constructs distortion compensation weights for each pixel based on Gladstone Dale constant and the ideal gas law. This effectively corrects the optical distortion caused by the change in air refractive index gradient due to surface thermal radiation during the hydrogen charging and discharging process of the hydrogen storage reactor, and eliminates the error of existing DIC algorithm in tracking pixel subsets in thermal field environments.

[0050] 2. Adaptive subset edge length matching improves the accuracy of deformation measurement in edge regions: By combining distortion compensation weights with the geometric curvature of the hydrogen storage reactor, an adaptive subset edge length is constructed. For pixels near the edge of the hydrogen storage reactor, the subset edge length is automatically increased to alleviate speckle aliasing and blurring caused by tangential perspective distortion and random refraction of the thermal boundary layer, significantly improving the robustness and spatial resolution of the full-field strain calculation.

[0051] 3. Constructing a stress concentration index for quantitative identification of hotspot regions: A stress concentration index is defined by utilizing the ratio of local stretching rate to global average stretching rate, the vector synthesis of stretching rate gradients within the neighborhood, and the physical scale corresponding to the subset side lengths. This index can effectively distinguish between uniform elastic expansion and localized abnormal stress concentration, eliminate interference from microscopic surface defects or image noise, and accurately identify macroscopic hotspot regions with a tendency for structural tearing.

[0052] 4. Dynamic early warning index to adapt to the sensitivity requirements of different pressurization stages: Based on the aspect ratio of the hydrogen storage reactor, real-time internal pressure, and rated permissible pressure, an early warning index that changes non-linearly with the pressurization process is constructed. The initial threshold is appropriately relaxed to avoid false alarms, and the threshold is automatically tightened as the pressure increases, improving the sensitivity to stress concentration in the later stages. This achieves an adaptive early warning mechanism that coordinates with the elasticity and plasticity of the container material and the safety margin.

[0053] 5. Achieving non-contact, real-time, early warning, and linked safety control: Based on a non-contact monitoring solution using a binocular stereo industrial camera and infrared thermal imager, there is no need to place sensors on the surface of the hydrogen storage reactor, avoiding installation risks under high temperature and high pressure environments. When the stress concentration index in a certain area is determined to exceed the dynamic warning index, the system immediately issues a warning message and automatically cuts off the power source of the hydrogen refueling unit and activates the pressure relief bypass valve, forming a complete closed-loop protection from deformation detection to emergency control, effectively preventing the formation of macroscopic cracks and hydrogen leakage accidents. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1 This is a flowchart illustrating the steps of a machine vision-based early warning method for stress concentration hotspots in a hydrogen storage reactor, according to an embodiment of the present invention. Detailed Implementation

[0056] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the machine vision-based early warning method and system for stress concentration hotspots in hydrogen storage reactors proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0057] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0058] Example 1:

[0059] This invention provides a machine vision-based method for early warning of stress concentration hotspots in hydrogen storage reactors, specifically as follows: Figure 1 As shown, it includes:

[0060] Step S001: Spray a speckle pattern onto the surface of the hydrogen storage reactor. When the hydrogen storage reactor begins to be filled with hydrogen, take pictures of the surface of the hydrogen storage reactor at fixed intervals to obtain multiple images.

[0061] Specifically, a high-contrast speckle pattern is sprayed onto the surface of the hydrogen storage reactor, and a binocular stereo industrial camera is positioned there. When the hydrogen storage reactor begins to be filled with hydrogen, images are taken of the reactor surface at fixed acquisition intervals to obtain multiple images. In this invention, a narrow-band filter is installed in front of the binocular stereo industrial camera lens; this is a high-frequency, short-wavelength blue light filter. Its purpose is to actively reduce the interference of infrared thermal radiation from high temperatures on the image acquisition process. Other filters can be installed in other embodiments. The preset acquisition interval in this embodiment is 15 seconds; other values ​​can be set according to user requirements in other embodiments. A smaller acquisition interval results in a greater computational load, but also stronger real-time warning capabilities. Furthermore, when photographing the hydrogen storage reactor, the optical axis of the binocular stereo industrial camera is aligned with the central axis of the reactor tank. Before hydrogen is filled, the internal pressure of the hydrogen storage reactor is less than [a certain value]. Furthermore, the method for obtaining the hydrogen storage reactor from the image is a well-known existing technology, and will not be described in detail in this embodiment.

[0062] This gives us all the images from each shot.

[0063] Step S002: Based on the pressure at the location of the hydrogen storage reactor when each image is captured, the real-time temperature of the area where the camera is located, the real-time temperature of the surface area of ​​the hydrogen storage reactor corresponding to each pixel in each speckle pattern in the image, the center wavelength of the filter installed in front of the camera lens, and the depth information of each pixel in the image, obtain the distortion compensation weight of each pixel in each speckle pattern in each image.

[0064] It should be noted that during rapid pressurization, the internal temperature of the hydrogen storage reactor rises sharply. This immense heat is conducted through the reactor walls to the outer surface, causing the outer wall temperature to be significantly higher than the temperature near the camera. This results in thermal flux disturbance, which in turn distorts the refractive index of the air around the reactor. Furthermore, the curvature of the reactor shell further exacerbates the distortion of the speckle pattern in areas with pronounced curvature. The algorithm is weakly resistant to light distortion. Therefore, the coupling between refractive index distortion and the curvature of the hydrogen storage reactor can cause errors in the algorithm's tracking of pixel subsets, leading to errors in the strain displacement results and rendering it unusable directly. The algorithm provides early warnings for the pressurization process of the hydrogen storage reactor. Therefore, optical distortion compensation weights are generated for each pixel within each speckle pattern in each acquired image.

[0065] It should be further noted that, since the refractive index gradient of air is the spatial derivative of air density, and the density distribution is significantly affected by the environment created by thermal radiation and convection on the surface of the hydrogen storage reactor, this invention selects a high-precision infrared thermal imager and an environmental temperature and pressure sensor positioned coaxially in the video monitoring system to collect relevant data. Based on this data, and using the classical optical refraction principle, an optical distortion compensation weight is generated for each pixel within each speckle pattern in the image. This weight value reflects the degree of targeted compensation required for the area where each pixel is located after being disturbed by thermal flow.

[0066] It should be further noted that the bending of light is related to the density difference between hot and cold air, and the rapid pressurization of the hydrogen storage reactor leads to a significant temperature difference between the surrounding air near the reactor and the ambient temperature in the camera's area. Furthermore, since the hydrogen storage reactor can be considered a heat source during pressurization, and its surface temperature differs from the normal ambient temperature, this scenario constitutes pure air convection heat transfer without chemical reactions. Therefore, the ideal gas law can be applied in this case. Calculate the air density at each location. In this equation... Indicates air density, This represents the specific gas constant of dry air. Indicates temperature. This represents air pressure. The derivation process for the density difference between the two regions is as follows: , The density difference between the two regions is This corresponds to the formula as follows: .in, This indicates the air density at lower temperatures. This indicates the temperature at a lower temperature location. This indicates the air density at higher temperatures. This indicates the temperature at a higher location. This indicates the temperature of the surface of the hydrogen storage reactor. This indicates the temperature of the camera's surrounding environment. This completes the calculation of the density difference.

[0067] It should be further explained that when obtaining the optical distortion compensation weight for each pixel in the image, since air density is directly proportional to the change in light refractive index, and Gladstone Dale's constant establishes the physical constitutive relationship between the fluid density field and the optical refractive index field, combining density differences with Gladstone Dale's constant, the thermodynamic dimension of fluid density differences is physically converted into the refractive index distortion rate in the optical dimension. This also reflects the influence of the temperature difference caused by the heat flow on the surface of the hydrogen storage reactor, which leads to air density differences, on the refractive index of light. Therefore, the optical distortion compensation weight for each pixel is calculated by combining density differences with Gladstone Dale's constant. Furthermore, the larger the Gladstone Dale's constant, the more sensitive the optical properties of the gaseous medium are to density changes.

[0068] It should be further noted that during image acquisition, a filter is typically installed in front of the camera lens to actively reduce the interference of infrared thermal radiation crosstalk caused by high temperatures on the image acquisition process. Under normal circumstances, this filter is usually a high-frequency, short-wavelength blue light filter. Furthermore, the longer the center wavelength of the filter, the more severe the background interference from the thermal radiation of the hydrogen storage reactor surface during acquisition. Longer wavelengths also have a greater optical diffraction limit, leading to a more blurred image background. Therefore, the longer the center wavelength of the filter, the greater the corresponding compensation weight should be. Thus, the optical distortion compensation weight for each pixel is calculated based on the center wavelength of the filter installed in front of the camera lens.

[0069] It should be further explained that since the heat emitted by the hydrogen storage reactor only exists within a certain range, it can be considered that a thermal flux boundary layer exists around the reactor. This thermal flux boundary layer has a thickness. Therefore, the smaller the distance between the camera and the region corresponding to a pixel in the image, the less interference this layer will cause in acquiring that pixel. This is because the smaller the distance, the shorter the length of the thermal flux boundary layer that light passes through during the imaging process. Therefore, the optical distortion compensation weight for each pixel is calculated based on the depth information of each pixel in the image.

[0070] Specifically, an infrared thermal imager array is deployed to acquire the real-time temperature of the surface area of ​​the hydrogen storage reactor corresponding to each pixel within each speckle pattern in the image when each image is captured, and to acquire the real-time temperature of the area where the binocular stereo industrial camera used for capturing each image is located. The method for acquiring the surface area of ​​the hydrogen storage reactor corresponding to each pixel within each speckle pattern in the image is a known prior art and will not be described in detail in this embodiment.

[0071] Furthermore, the atmospheric pressure sensor located at the site of the hydrogen storage reactor is used to obtain the pressure at the location of the hydrogen storage reactor when each image is taken. The pixels within each speckle pattern in each image are also obtained; this is existing known technology and will not be described in detail in this embodiment.

[0072] Get the The first image The first speckle pattern within the first The specific formula for calculating the distortion compensation weight of each pixel is as follows:

[0073]

[0074] In the formula, Indicates the first The first image The first speckle pattern within the first Distortion compensation weights for each pixel Gladstonedale constant represents the relationship between the polarizability of air molecules and the macroscopic refractive index. (Unit: [not specified]) ; Indicates the shooting of the first The image shows the pressure at the location of the hydrogen storage reactor, in units of... , equivalent to ; The specific gas constant of dry air, in units of , equivalent to ; Indicates the shooting of the first The image shows the real-time temperature of the area where the stereo industrial camera used for taking the picture is located, in Kelvin. , Indicates the shooting of the first The first image The first image The first speckle pattern within the first Each pixel corresponds to the real-time temperature of the surface region of the hydrogen storage reactor, in Kelvin. , Represents the absolute value function. This indicates the center wavelength of the narrowband filter mounted in front of the lens of a stereo industrial camera used for shooting, in units of... ; Indicates the first The first image The first speckle pattern within the first The depth information of each pixel, in meters, reflects the relationship between the stereo industrial camera used for shooting and the first... The first image The first speckle pattern within the first The distance between the regions corresponding to each pixel. When acquiring images using a binocular stereo industrial camera, each pixel in the image has depth information; the depth information of each pixel in the image is not zero, and acquiring the depth information of pixels in the image is a known existing technique. , , The methods described are all existing well-known technologies, and will not be elaborated upon in this embodiment. Since 0 Kelvin equals -273.15 degrees Celsius, ... Not zero.

[0075] It should be noted that, Indicates the first The first image The first speckle pattern within the first The air density around the area corresponding to each pixel is related to the image capture. The difference in air density in the area where the stereo industrial camera used for image capture is located; wherein, in this embodiment, the temperature is measured in Kelvin, making... and It cannot be 0; through and Multiplication converts the difference in fluid density in the thermodynamic dimension into the refractive index distortion rate in the optical dimension; The larger the value, the more severe the background interference from the thermal radiation on the surface of the hydrogen storage reactor during image capture, and the greater the optical diffraction limit at longer wavelengths, leading to a more blurred image background. Therefore, by... calculate ,and The larger the value, the greater the compensation that should be received.

[0076] It should be further explained that, since the heat dissipated from the hydrogen storage reactor to the surroundings exists only within a certain range, it can be considered that a heat flux boundary layer exists around the hydrogen storage reactor. This heat flux boundary layer has a thickness, which makes it possible for the heat to dissipate from the reactor to increase in temperature and humidity. The smaller the value, the less influence the thermal flow boundary layer has on the refraction of light during the shooting process, because during the shooting process, the thermal flow boundary layer... The first image The first speckle pattern within the first When light emitted from the region corresponding to each pixel reaches the camera, the thinner the thermal boundary layer it passes through, the less interference it experiences.

[0077] Thus, the first... The first image The first speckle pattern within the first Distortion compensation weights for each pixel.

[0078] Step S003: Based on the distortion compensation weight of each pixel within each speckle pattern in each image, its position within the image, and using... The algorithm initializes the subset side length of each pixel and obtains the adaptive subset side length of each pixel within each speckle pattern in each image.

[0079] It should be noted that the use When matching pixels in two images, the algorithm needs to use a subset of the side lengths of each pixel. This is because different pixels within the same image may experience different distortions during image capture. Therefore, using… When implementing the algorithm, it is necessary to adaptively adjust the side length of a subset of each pixel in the image.

[0080] It's important to further explain that the hydrogen storage reactor itself can be considered a typical large-curvature cylinder. When the optical axis of the industrial camera is directly aligned with the central axis of the reactor tank, the normal vector of the reactor surface is parallel to the optical axis, resulting in an orthographic projection with normal speckle characteristics. However, as the observed point extends towards the edges of the cylinder, the angle between the normal vector of the observed point's surface and the camera's optical axis increases dramatically. This causes significant tangential perspective distortion, where a very small number of pixels at the edges correspond to surface information of the tank that should have a much larger number of pixels. This leads to spatial aliasing or even severe blurring in the speckle area. Simultaneously, when the camera is directly aligned with the central axis, light rays tend to travel a shorter straight-line distance along the normal vector direction through the thermal flux boundary layer formed on the reactor tank surface. However, at the edges, light rays travel along the tangent direction of the curved surface through the thermal flux layer. During this process, the accumulation of random refraction distortion is significantly more severe. Therefore, as mentioned above, the closer to the edge region, the more pronounced the tendency for feature loss and refraction distortion becomes. Therefore, an adaptive subset side length is defined for pixels at different locations in the image.

[0081] It should be further noted that a larger distortion compensation weight for a pixel indicates a stronger thermal flux disturbance experienced by the corresponding photosensitive unit of the camera when obtaining information about the pixel's location. In this case, the subset side length of that pixel is enlarged. Therefore, the optical distortion compensation weight of each pixel is used when determining the adaptive subset side length.

[0082] It should be further explained that when determining the adaptive subset side length for each pixel, the closer a pixel is to the central axis of the hydrogen storage reactor, the smaller the distance between the pixel and the central sectional plane of the hydrogen storage reactor and the cross-section of the reactor at its location. In this case, the interference encountered when acquiring the pixel is less. Therefore, the adaptive subset side length for each pixel is calculated based on its position within the image. The more interference a pixel experiences, the larger its adaptive subset side length should be. This is because when using the DIC algorithm, a larger subset side length provides stronger noise resistance.

[0083] Specifically, the first Using the centroid of the hydrogen storage reactor in the image as the origin, a two-dimensional coordinate system is constructed with a straight line parallel to the central axis of the reactor tank as the vertical axis and a straight line perpendicular to the vertical axis as the horizontal axis. The method for obtaining the centroid of a region in the image is a known prior art and will not be described in detail in this embodiment. Similarly, the method for obtaining the central axis of the hydrogen storage reactor tank is also a known prior art and will not be described in detail in this embodiment.

[0084] Furthermore, the first The first image The first speckle pattern within the first The value of the nth pixel on the horizontal axis is denoted as the nth pixel. The first image The first speckle pattern within the first The distance between each pixel and the central axis of the hydrogen storage reactor.

[0085] Furthermore, obtain the first The first image The first speckle pattern within the first The specific formula for calculating the side length of the adaptive subset of each pixel is as follows:

[0086]

[0087] In the formula, Indicates the first The first image The first speckle pattern within the first Adaptive subset side length of each pixel Indicates use The algorithm initializes the side length of a subset of pixels. Indicates the first The first image The first speckle pattern within the first Distortion compensation weights for each pixel This indicates the radius of the cylindrical section of the hydrogen storage reactor when it is not filled with gas, in meters. This represents the physical spatial resolution parameter of the stereo industrial camera used for imaging, after calibration on the front surface of the hydrogen storage reactor. It is the length of the smallest line segment within the physical plane corresponding to a single pixel in the image, expressed in units of... ; Indicates the first The first image The first speckle pattern within the first The distance of each pixel to the central axis of the hydrogen storage reactor. And obtain the data using... The initial values ​​for the subset side lengths of pixels in the algorithm are existing well-known techniques and will not be elaborated upon in this embodiment. Furthermore, it uses... The initial value of the subset side length of the pixel in the algorithm refers to, not... The algorithm was improved and used directly. In the algorithm, the side length of a subset of pixels is used. Obtaining the radius of the cylindrical section of the hydrogen storage reactor when it is not filled with gas, obtaining the physical spatial resolution parameters of the camera after calibration on the object surface, and obtaining the radius of the cylindrical section of the hydrogen storage reactor when it is not filled with gas are all existing well-known technologies, and will not be elaborated upon in this embodiment. Furthermore, through 1 and... When adding, 1 and The units are the same. Among them, obtaining... The method is a well-known existing technology, and will not be described in detail in this embodiment.

[0088] It should be noted that, This reflects the first The first image The first speckle pattern within the first The distance between each pixel and the vertical axis This reflects the mapping of the vertical axis in the image to the tank body, the first... The first image The first speckle pattern within the first The actual distance between each pixel and the vertical axis Adding 1 is to prevent the denominator from being 0; The larger the value, the more obvious the observation point's position on the edge of the column; in this case, the more important it is to increase the value of the second [value]. The first image The first speckle pattern within the first The side length of a subset of pixels. And The result is greater than or equal to 0.

[0089] Thus, the adaptive subset side length of each pixel within each speckle pattern in each image is obtained.

[0090] Step S004: Based on the adaptive subset side length of each pixel in each speckle pattern in each image, obtain the stretching ratio of each pixel in each speckle pattern in each image; based on the stretching ratio of each pixel in each speckle pattern in each image, the adaptive subset side length, and the stretching ratio of surrounding pixels, obtain the stress concentration index of each pixel in each speckle pattern in each image.

[0091] It should be noted that the adaptive subset side length of each pixel in the image is used for... During the matching process, the algorithm uses the image of the hydrogen storage reactor in an absolutely static state before it is filled with hydrogen as the constant matching object. When hydrogen filling begins, each image collected during the hydrogen filling process is matched with the image before hydrogen filling. During the matching process, the algorithm obtains the corresponding search range of each pixel based on the subset side length of each pixel, thereby obtaining the displacement of each pixel under the comparison of the two images, and thus obtaining the stretching rate of each pixel.

[0092] It should be further noted that, in the current scenario, when the hydrogen storage reactor is filled with ultra-high-pressure hydrogen, the entire outer shell of the reactor will undergo elastic expansion. This expansion will cause all pixels on the surface of the reactor to undergo a certain degree of displacement and stretching. However, the degree of impact of expansion will definitely differ between areas of stress concentration and areas with gentle and uniform stress distribution. Therefore, the stress concentration index of each pixel needs to be calculated.

[0093] It should be further noted that when calculating the stress concentration index for each pixel, if the current pixel is simply expanding normally along with the entire hydrogen storage reactor during gasification, then the deformation at that pixel location will be synchronized with the global average deformation. Therefore, the stress concentration index for each pixel in the image is calculated by comparing the stretching rate of each pixel in each image with the average stretching rate of all pixels in all images.

[0094] It should be further noted that the expansion effect experienced by pixels in abnormal and normal regions will inevitably differ. This difference will lead to a situation where, when a pixel is in a region with abnormal stress, the stretching ratio of other pixels around that pixel will also be relatively large. Therefore, the stress concentration index of each pixel in the image is calculated based on the distribution of the stretching ratio of other pixels around each pixel.

[0095] It should be further explained that for each pixel, During algorithm matching, if a pixel still exhibits a relatively high stretching rate when observed at a larger scale, then physically, this indicates that the anomaly is not caused by microscopic surface defects or isolated image noise, but rather tends to indicate a more macroscopic structural tearing occurring over a wider area. That is, the larger the physical span affected by the anomaly, the greater the tendency for locally accumulated destructive energy, and the higher the corresponding stress concentration index. Therefore, according to... The algorithm calculates the stress concentration index for each pixel in the image during each matching process.

[0096] Specifically, using the DIC algorithm, based on the first... The first image The first speckle pattern within the first The adaptive subset side length of the nth pixel is obtained to get the nth pixel. The first image The first speckle pattern within the first The displacement value of the nth pixel, according to the nth The displacement values ​​of all pixels in the image are used to obtain the first... The first image The first speckle pattern within the first The stretching rate of each pixel is described. The DIC algorithm is used to obtain the displacement value of each pixel in each image, which is a known technique and will not be elaborated upon in this embodiment. Furthermore, the stretching rate of the pixel is obtained based on the displacement value, which is also a known technique and will not be elaborated upon in this embodiment. When using the DIC algorithm, if it is necessary to use the stretching rate of the first pixel... Pixels in the image that are not located within the speckle pattern are then used. The method of obtaining the first The adaptive subset side length of pixels not located within the speckle pattern in an image.

[0097] Furthermore, the position located at the The first image The first speckle pattern within the first The pixel within the eight-neighborhood of the i-th pixel, and located within the speckle pattern, is denoted as the i-th pixel. The first image The first speckle pattern within the first The neighboring pixels of a pixel.

[0098] Furthermore, obtain the first The first image The first speckle pattern within the first The specific formula for calculating the stress concentration index of each pixel is as follows:

[0099]

[0100] In the formula, Indicates the first The first image The first speckle pattern within the first Stress concentration index per pixel Indicates the first The mean stretching ratio of pixels within all speckle patterns in the image. Indicates the first The first image The first speckle pattern within the first Strength per pixel Indicates the first The first image The first speckle pattern within the first The average stretching ratio of all neighboring pixels of a given pixel. Indicates the first The first image The first speckle pattern within the first Adaptive subset side length of each pixel This represents the physical spatial resolution parameters of the binocular stereo industrial camera used for filming, after calibration on the front surface of the hydrogen storage reactor. It is an exponential function with the natural constant as its base. It represents 1 meter.

[0101] It should be noted that, for Part, if the first The first image The first speckle pattern within the first If a pixel expands during normal gasification of the entire hydrogen storage reactor, the deformation at that pixel location will be synchronized with the global average deformation, and the corresponding ratio will be close to... However, if the stress at that point is abnormal, it will lead to Significantly greater than or equal to This leads to a larger overall value for that part, and at this time the corresponding It should also be larger; for part, This reflects the stretching ratio of other pixels surrounding the given pixel. Because when this pixel is in an area with abnormal stress, the stretching ratio of other pixels around it is also relatively large, making... The larger the value, the more pronounced the tendency for abnormal stress in the region where the pixel is located; using The function is used because the risk of structural failure due to stress concentration does not increase linearly with the elongation rate. Instead, the risk increases as the physical extent of the local deformation expands and the neighborhood anomaly becomes more pronounced. This physical law of function matching allows the index to more accurately reflect the risk level; through Used to eliminate dimensions; due to The unit can be considered as the number of pixels, making... The unit is meters.

[0102] It should be further explained that, Part of it reflects the scale range in which pixels can be reliably matched. If a large stretching rate is observed in some pixels at a larger scale, then in a physical sense, it means that the anomaly is not caused by microscopic surface defects or a few isolated image noises, but rather tends to be a more macroscopic structural tearing that is occurring in a wider area. That is, the larger the physical span affected by the anomaly, the greater the tendency of locally accumulated destructive energy, and therefore the greater the corresponding stress concentration index.

[0103] Thus, the stress concentration index of each pixel within each speckle pattern in each image is obtained.

[0104] Step S005: Based on the pressure inside the hydrogen storage reactor, the shape of the hydrogen storage reactor, and the stress concentration index of each pixel in each speckle pattern in each image when each image is acquired, an early warning is given for the gasification process of the hydrogen storage reactor.

[0105] It should be noted that initially, the hydrogen storage reactor has ample overall resistance during the initial filling process. If minor imperfections cause slight stress concentration on the reactor surface, the reactor material's elasticity and plasticity can completely absorb the stress, and the threshold should be appropriately increased. However, as the filling process continues, the pressure in the reactor gradually increases, leading to increased stress. Furthermore, the reactor's sensitivity to stress increases, and the threshold should be decreased to prevent safety hazards caused by any localized stress concentration. Therefore, the warning index is adjusted as the filling time progresses.

[0106] Specifically, the internal pressure of the hydrogen storage reactor before it is filled with hydrogen is recorded as the initial pressure. When hydrogen is filled into the hydrogen storage reactor, its internal pressure rises from the initial pressure to... At that time, images were acquired using the method in step S001. When the hydrogen storage reactor was filled with hydrogen, the internal pressure rose to... The mean stress concentration index of all pixels within all speckle patterns in all images acquired at that time is recorded as the stress concentration index of the internal pressure of the hydrogen storage reactor rising from the initial pressure to the value obtained when hydrogen is introduced into the reactor. The average stress concentration index of the entire tank of the hydrogen storage reactor. The stress concentration index of the pixels in this section is obtained using the method in step S004. As a benchmark value because This represents the baseline stress level of a hydrogen storage reactor during the standard pressurization process. It is an inherent reference value strongly correlated with the reactor's structure and material properties. Multiplying this value by a dynamic adjustment coefficient allows the warning index to be adapted to the reactor's inherent properties and flexibly adjusted according to real-time pressure changes, avoiding the problem of a universal threshold not matching different reactor specifications. 10 MPa was chosen because it is a benchmark test pressure point during the pressurization process of a hydrogen storage reactor; 10 MPa is a common performance test node for high-pressure hydrogen storage reactors. At this stress level, the overall deformation characteristics of the vessel are already reflected, and it is far below the ultimate pressure, making the data stable and reliable, suitable as a benchmark reference. Therefore, [the following text is incomplete and requires further context: "…"] As a benchmark, this invention does not cause the internal pressure of the hydrogen storage reactor to rise from the initial pressure when hydrogen is introduced into the reactor. Early warnings were issued.

[0107] Furthermore, obtain the first The specific formula for calculating the warning index of Zhang image is as follows:

[0108]

[0109] In the formula, Indicates the first The warning index of the image, This indicates that when hydrogen is introduced into the hydrogen storage reactor, the internal pressure rises from the initial pressure to... The average stress concentration index of the entire tank of the hydrogen storage reactor. This indicates the design axial length of the tank in the hydrogen storage reactor, in units of... , This represents the radius of the cylindrical section of the hydrogen storage reactor when it is not filled with gas, in units of... , This indicates the rated permissible pressure for the hydrogen storage reactor to operate at its limit, in units of... , Indicates obtaining the first The image shows the pressure inside the hydrogen storage reactor, in units of [value missing]. , The value is a logarithmic function with the natural constant as its base. The radius of the cylindrical section of the hydrogen storage reactor when it is not filled with gas, the designed axial length of the reactor tank, and the rated permissible pressure for extreme operation of the hydrogen storage reactor are all existing well-known technologies and will not be elaborated upon in this embodiment. This represents a natural constant with a value of approximately 2.71828.

[0110] It should be noted that, The length-to-diameter ratio represents the ratio of the major axis length to the radius of the cylindrical section of the hydrogen storage reactor. A larger length-to-diameter ratio indicates a more sensitive distribution of overall deformation energy, and this ratio is used as a correction parameter in the calculation. Part of, for Partly, at the beginning Smaller, as the collection time increases As the value gradually increases, the overall value of this part gradually decreases. The magnification also gradually decreases, and this gradual decrease will cause... The function result gradually approaches and use The function reflects that the sensitivity to early warning increases as the pressurization process progresses. add It is used to prevent The function outputs numbers less than 1; use The ln function is used because in the initial stage of inflation, the internal pressure is low, and the container material has a large elastic-plastic margin. Even if small stress concentrations occur, they can be absorbed by the material, requiring a lower warning sensitivity to avoid false alarms. In the later stage of inflation, the pressure approaches the rated permissible pressure. At this point, the result of the ln function gradually approaches 1, and the corresponding threshold tightens, significantly increasing the sensitivity to stress concentration, which meets the actual needs of increased safety risks in the later stages. Furthermore, the growth and decay of the ln function are smooth and non-linear, unlike linear functions which show sharp increases and decreases in the threshold with pressure changes, making the warning logic more stable.

[0111] Furthermore, when Then it is determined that in the collection of the first When the image is displayed, the first... The first image The first speckle pattern within the first When a pixel is located in an area with local microscopic defects and structural damage, resulting in abnormal stress deformation, an early warning is issued. At the same time, the warning information is reported to the terminal, and the central relay uses the Ethernet protocol to cut off the power supply to the high-power high-pressure scroll compressor at the front end of the hydrogen refueling unit and urgently starts the pressure relief and discharge bypass solenoid valve to release the pressure.

[0112] This concludes the embodiment.

[0113] A second aspect of the present invention provides a machine vision-based early warning system for stress concentration hotspots in a hydrogen storage reactor. This system includes a data acquisition module, a distortion compensation weight calculation module, an adaptive subset side length module, a stress concentration index calculation module, and an early warning module, wherein:

[0114] The data acquisition module is used to spray a speckle pattern on the surface of the hydrogen storage reactor. When the hydrogen storage reactor starts to be filled with hydrogen, the surface of the hydrogen storage reactor is photographed at fixed intervals to obtain multiple images.

[0115] The distortion compensation weight calculation module is used to obtain the distortion compensation weight of each pixel in each speckle pattern in each image based on the pressure of the hydrogen storage reactor when each image is captured, the real-time temperature of the area where the camera is located, the real-time temperature of the surface area of ​​the hydrogen storage reactor corresponding to each pixel in each speckle pattern in the image, the center wavelength of the filter installed in front of the camera lens, and the depth information of each pixel in the image.

[0116] The adaptive subset side length module is used to determine the distortion compensation weights of each pixel within each speckle pattern in each image, its position within the image, and the use of... The algorithm initializes the subset side length of each pixel and obtains the adaptive subset side length of each pixel within each speckle pattern in each image.

[0117] The stress concentration index calculation module is used to obtain the stretching ratio of each pixel in each speckle pattern in each image based on the adaptive subset side length of each pixel in each speckle pattern in each image; and to obtain the stress concentration index of each pixel in each speckle pattern in each image based on the stretching ratio of each pixel in each speckle pattern in each image, the adaptive subset side length and the stretching ratio of the surrounding pixels.

[0118] The early warning module is used to provide early warnings about the gasification process of the hydrogen storage reactor based on the pressure inside the hydrogen storage reactor, the shape of the hydrogen storage reactor, and the stress concentration index of each pixel in each speckle pattern in each image when each image is acquired.

Claims

1. A machine vision-based method for early warning of stress concentration hotspots in hydrogen storage reactors, characterized in that, The method includes the following steps: A speckle pattern was sprayed onto the surface of the hydrogen storage reactor. When the hydrogen storage reactor began to be filled with hydrogen, the surface of the hydrogen storage reactor was photographed at fixed intervals to obtain multiple images. Based on the pressure at the location of the hydrogen storage reactor when each image is captured, the real-time temperature of the area where the camera is located, the real-time temperature of the surface area of ​​the hydrogen storage reactor corresponding to each pixel in each speckle pattern in the image, the center wavelength of the filter installed in front of the camera lens, and the depth information of each pixel in the image, the distortion compensation weight of each pixel in each speckle pattern in each image is obtained. Based on the distortion compensation weights of each pixel within each speckle pattern in each image, its position within the image, and using... The algorithm initializes the subset side length of each pixel and obtains the adaptive subset side length of each pixel within each speckle pattern in each image. Based on the adaptive subset side length of each pixel in each speckle pattern in each image, the stretching ratio of each pixel in each speckle pattern in each image is obtained; based on the stretching ratio of each pixel in each speckle pattern in each image, the adaptive subset side length and the stretching ratio of the surrounding pixels, the stress concentration index of each pixel in each speckle pattern in each image is obtained. The gasification process of the hydrogen storage reactor is monitored based on the internal pressure and shape of the reactor when each image is acquired, as well as the stress concentration index of each pixel within each speckle pattern in each image.

2. The machine vision-based early warning method for stress concentration hotspots in a hydrogen storage reactor according to claim 1, characterized in that, The specific calculation steps for spraying a speckle pattern onto the surface of the hydrogen storage reactor and taking multiple images of the reactor surface at fixed intervals when hydrogen is started to be introduced into the reactor are as follows: A high-contrast speckle pattern is sprayed onto the surface of the hydrogen storage reactor. A binocular stereo industrial camera is set up with a set acquisition interval. When the hydrogen storage reactor starts to be filled with hydrogen, the surface of the hydrogen storage reactor is photographed at the set acquisition interval to obtain multiple images. When photographing the hydrogen storage reactor, the optical axis of the binocular stereo industrial camera is aligned with the central axis of the hydrogen storage reactor tank.

3. The machine vision-based early warning method for stress concentration hotspots in a hydrogen storage reactor according to claim 1, characterized in that, The specific calculation steps for obtaining the distortion compensation weight of each pixel in each speckle pattern in each image, based on the pressure at the location of the hydrogen storage reactor when each image is captured, the real-time temperature of the area where the camera is located, the real-time temperature of the surface area of ​​the hydrogen storage reactor corresponding to each pixel in each speckle pattern in the image, the center wavelength of the filter installed in front of the camera lens, and the depth information of each pixel in the image, are as follows: An infrared thermal imager array is deployed to obtain the real-time temperature of the surface area of ​​the hydrogen storage reactor corresponding to each pixel in each speckle pattern in the image when each image is captured, and to obtain the real-time temperature of the area where the binocular stereo industrial camera used for capturing the image is located when each image is captured. The pressure at the location of the hydrogen storage reactor was obtained by using atmospheric pressure sensors placed at the site of the hydrogen storage reactor when each image was taken. Get the The first image The first speckle pattern within the first The specific formula for calculating the distortion compensation weight of each pixel is as follows: In the formula, Indicates the first The first image The first speckle pattern within the first Distortion compensation weights for each pixel Gladstone constant representing air. Indicates the shooting of the first The image shows the pressure at the location of the hydrogen storage reactor. This represents the specific gas constant of dry air. Indicates the shooting of the first The image shows the real-time temperature of the area where the binocular stereo industrial camera used for taking the picture. Indicates the shooting of the first The first image The first image The first speckle pattern within the first Each pixel corresponds to the real-time temperature of the surface area of ​​the hydrogen storage reactor. Represents the absolute value function. This indicates the center wavelength of the narrowband filter installed in front of the lens of a stereo industrial camera used for shooting. Indicates the first The first image The first speckle pattern within the first Depth information of each pixel.

4. The machine vision-based early warning method for stress concentration hotspots in a hydrogen storage reactor according to claim 1, characterized in that, The distortion compensation weights for each pixel within each speckle pattern in each image, their positions within the image, and the use of... The algorithm initializes the subset side length of pixels and calculates the adaptive subset side length of each pixel within each speckle pattern in each image using the following steps: The first Using the centroid of the hydrogen storage reactor in the image as the origin, a two-dimensional coordinate system is constructed with a straight line parallel to the central axis of the hydrogen storage reactor tank as the vertical axis and a straight line perpendicular to the vertical axis as the horizontal axis. The first The first image The first speckle pattern within the first The value of the nth pixel on the horizontal axis is denoted as the nth pixel. The first image The first speckle pattern within the first The distance between each pixel and the central axis of the hydrogen storage reactor; Get and use The initial values ​​of the subset side lengths of pixels in the algorithm, and the radius of the cylindrical section of the hydrogen storage reactor when it is not filled with gas; According to usage The algorithm initializes the side length of the subset of pixels and the first... The first image The first speckle pattern within the first The distortion compensation weights of each pixel, the distance from the central axis of the hydrogen storage reactor, the radius of the cylindrical section of the hydrogen storage reactor when it is not filled with gas, and the physical spatial resolution parameters of the binocular stereo industrial camera used for shooting after calibration on the front surface of the hydrogen storage reactor are used to obtain the first pixel. The first image The first speckle pattern within the first Adaptive subset side length for each pixel.

5. The machine vision-based early warning method for stress concentration hotspots in a hydrogen storage reactor according to claim 4, characterized in that, According to the use The algorithm initializes the side length of the subset of pixels and the first... The first image The first speckle pattern within the first The distortion compensation weights of each pixel, the distance from the central axis of the hydrogen storage reactor, the radius of the cylindrical section of the hydrogen storage reactor when it is not filled with gas, and the physical spatial resolution parameters of the binocular stereo industrial camera used for shooting after calibration on the front surface of the hydrogen storage reactor are used to obtain the first pixel. The first image The first speckle pattern within the first The specific steps for calculating the adaptive subset side length of each pixel are as follows: In the formula, Indicates the first The first image The first speckle pattern within the first Adaptive subset side length of each pixel Indicates use The algorithm initializes the side length of a subset of pixels. Indicates the first The first image The first speckle pattern within the first Distortion compensation weights for each pixel This indicates the radius of the cylindrical section of the hydrogen storage reactor when it is not filled with gas. This represents the physical spatial resolution parameters of the binocular stereo industrial camera used for filming, after calibration on the front surface of the hydrogen storage reactor. Indicates the first The first image The first speckle pattern within the first The distance between each pixel and the central axis of the hydrogen storage reactor.

6. The machine vision-based early warning method for stress concentration hotspots in a hydrogen storage reactor according to claim 1, characterized in that, The specific calculation steps for obtaining the stretching ratio of each pixel in each speckle pattern of each image based on the adaptive subset side length of each pixel in each speckle pattern of each image are as follows: Using the DIC algorithm, according to the first... The first image The first speckle pattern within the first The adaptive subset side length of the nth pixel is obtained to get the nth pixel. The first image The first speckle pattern within the first The displacement value of the nth pixel, according to the nth The displacement values ​​of all pixels in the image are used to obtain the first... The first image The first speckle pattern within the first Stretching rate per pixel.

7. The machine vision-based early warning method for stress concentration hotspots in a hydrogen storage reactor according to claim 1, characterized in that, The specific calculation steps for obtaining the stress concentration index of each pixel in each speckle pattern of each image based on the stretching ratio of each pixel in each speckle pattern of each image, the adaptive subset side length, and the stretching ratio of surrounding pixels are as follows: Will be located in the The first image The first speckle pattern within the first The pixel within the eight-neighborhood of the i-th pixel, and located within the speckle pattern, is denoted as the i-th pixel. The first image The first speckle pattern within the first The neighboring pixels of a pixel; Get the The first image The first speckle pattern within the first The specific formula for calculating the stress concentration index of each pixel is as follows: In the formula, Indicates the first The first image The first speckle pattern within the first Stress concentration index per pixel Indicates the first The mean stretching ratio of pixels within all speckle patterns in the image. Indicates the first The first image The first speckle pattern within the first Strength per pixel Indicates the first The first image The first speckle pattern within the first The average stretching ratio of all neighboring pixels of a given pixel. Indicates the first The first image The first speckle pattern within the first Adaptive subset side length of each pixel This represents the physical spatial resolution parameters of the binocular stereo industrial camera used for filming, after calibration on the front surface of the hydrogen storage reactor. It is an exponential function with the natural constant as its base. It represents 1 meter.

8. The machine vision-based early warning method for stress concentration hotspots in a hydrogen storage reactor according to claim 1, characterized in that, The specific calculation steps for providing early warning of the hydrogen storage reactor's charging process based on the internal pressure of the hydrogen storage reactor, the shape of the hydrogen storage reactor, and the stress concentration index of each pixel within each speckle pattern in each image are as follows: Obtain the internal pressure of the hydrogen storage reactor before it is filled with hydrogen, and record it as the initial pressure. The internal pressure of the hydrogen storage reactor increased from the initial pressure when hydrogen was introduced into the reactor. The mean stress concentration index of all pixels within all speckle patterns in all images acquired at that time is recorded as the stress concentration index of the internal pressure of the hydrogen storage reactor rising from the initial pressure to the value obtained when hydrogen is introduced into the reactor. The average stress concentration index of the entire tank of the hydrogen storage reactor. ; Based on the pressure inside the hydrogen storage reactor when each image was acquired, the shape of the hydrogen storage reactor, and the increase in internal pressure from the initial pressure to the pressure when hydrogen was introduced into the hydrogen storage reactor. The stress concentration index of the entire tank of the hydrogen storage reactor was obtained. The warning index of the image; If the first In the image, the stress concentration index of pixels located within the speckle pattern is greater than that of the first pixel. The warning index of the first image determines the warning level at the time of collection. When the image is captured, abnormal stress deformation caused by local microscopic defects and structural damage on the surface of the hydrogen storage reactor is detected. At this time, the early warning program is triggered and a warning is reported to the terminal.

9. The machine vision-based early warning method for stress concentration hotspots in a hydrogen storage reactor according to claim 8, characterized in that, The process is based on the pressure inside the hydrogen storage reactor when each image is acquired, the shape of the hydrogen storage reactor, and the increase in internal pressure from the initial pressure when hydrogen is introduced into the hydrogen storage reactor. The stress concentration index of the entire tank of the hydrogen storage reactor was obtained. The specific steps for calculating the warning index of Zhang image are as follows: In the formula, Indicates the first The warning index of the image, This indicates that when hydrogen is introduced into the hydrogen storage reactor, the internal pressure rises from the initial pressure to... The average stress concentration index of the entire tank of the hydrogen storage reactor. This indicates the design axis length of the tank body of the hydrogen storage reactor. This indicates the radius of the cylindrical section of the hydrogen storage reactor when it is not filled with gas. This indicates the rated permissible pressure for the hydrogen storage reactor to operate at its limit. Indicates obtaining the first The pressure inside the hydrogen storage reactor when the image is taken. Let be a logarithmic function with the natural constant as its base. It is a natural constant.

10. A machine vision-based early warning system for stress concentration hotspots in a hydrogen storage reactor, characterized in that, The system includes the following modules: The data acquisition module is used to spray a speckle pattern on the surface of the hydrogen storage reactor. When the hydrogen storage reactor starts to be filled with hydrogen, the surface of the hydrogen storage reactor is photographed at fixed intervals to obtain multiple images. The distortion compensation weight calculation module is used to obtain the distortion compensation weight of each pixel in each speckle pattern in each image based on the pressure of the hydrogen storage reactor when each image is captured, the real-time temperature of the area where the camera is located, the real-time temperature of the surface area of ​​the hydrogen storage reactor corresponding to each pixel in each speckle pattern in the image, the center wavelength of the filter installed in front of the camera lens, and the depth information of each pixel in the image. The adaptive subset side length module is used to determine the distortion compensation weights of each pixel within each speckle pattern in each image, its position within the image, and the use of... The algorithm initializes the subset side length of each pixel and obtains the adaptive subset side length of each pixel within each speckle pattern in each image. The stress concentration index calculation module is used to obtain the stretching ratio of each pixel in each speckle pattern in each image based on the adaptive subset side length of each pixel in each speckle pattern in each image; and to obtain the stress concentration index of each pixel in each speckle pattern in each image based on the stretching ratio of each pixel in each speckle pattern in each image, the adaptive subset side length and the stretching ratio of the surrounding pixels. The early warning module is used to provide early warnings about the gasification process of the hydrogen storage reactor based on the pressure inside the hydrogen storage reactor, the shape of the hydrogen storage reactor, and the stress concentration index of each pixel in each speckle pattern in each image when each image is acquired.

Images