A method and system for observing temperature distribution of a hydrogen-oxygen flame based on a flame fusion method single crystal growth

Abstract

The application discloses a hydrogen-oxygen flame temperature distribution observation method and system based on a flame fusion method single crystal growth, relates to the technical field of the flame fusion method single crystal growth, and comprises the following steps: S1, a double-waveband high-speed imaging device is installed at an observation hole of a flame fusion method single crystal growth device; two narrow-band filters are arranged on the double-waveband high-speed imaging device; a lens is aligned with a hydrogen-oxygen flame area; and hydrogen-oxygen flame image data of the hydrogen-oxygen flame at two wavebands is collected; S2, the hydrogen-oxygen flame image data is converted into digital signal data; a 3*3 convolution kernel Gaussian filter model is used to denoise the digital signal data; a polynomial distortion correction model is used to perform distortion correction processing on the denoised digital signal data; and a corrected gray value matrix is obtained. The application adopts the double-waveband high-speed imaging device to collect hydrogen-oxygen flame global image data, and solves the problem that a traditional contact type temperature measurement can only detect local point temperature by combining with pixel level temperature calculation, so that the distribution characteristics of a flame temperature field can be completely reflected.

Description

Technical Field

[0001] This invention relates to the field of flame fusion single crystal growth technology, and in particular to a method and system for observing the temperature distribution of an oxyhydrogen flame based on flame fusion single crystal growth. Background Technology

[0002] Flame fusion is an important method for the artificial synthesis of single crystals. The hydrogen-oxygen flame, as the core heat source for single crystal growth in flame fusion, directly determines the growth quality, crystal structure integrity, and performance indicators of the single crystal due to the different temperatures and stability of its distribution at various points.

[0003] Traditional methods for detecting the temperature of oxyhydrogen flames mostly employ contact thermometers, which can only measure the temperature at local points in the flame and cannot reflect the overall temperature distribution. Non-contact temperature measurement methods, such as ordinary monochromatic temperature measurement, are easily affected by factors such as flame emissivity and environmental interference, resulting in low measurement accuracy and difficulty in achieving high-speed, dynamic temperature field observation. At the same time, the existing temperature measurement methods have imperfect verification and calibration mechanisms, which cannot guarantee the reliability and consistency of the observation data. Summary of the Invention

[0004] The embodiments of the present invention provide a method and system for observing the temperature distribution of hydrogen-oxygen flame based on single crystal growth by flame melting, aiming to solve the problems of being unable to reflect the characteristics of the temperature distribution across the entire range, being unable to achieve high-speed and dynamic temperature field observation, and being unable to guarantee the reliability and consistency of the observation data.

[0005] To achieve the above objectives, the embodiments of the present invention adopt the following technical solutions:

[0006] A method and system for observing the temperature distribution of an oxyhydrogen flame based on flame melting single crystal growth includes the following steps:

[0007] S1: Install the dual-band high-speed imaging device at the observation hole of the flame melting single crystal growth device. The dual-band high-speed imaging device is equipped with two narrowband filters in the dual-band. The lens is aimed at the oxyhydrogen flame area to collect oxyhydrogen flame image data in the two bands.

[0008] S2: Convert the hydrogen-oxygen flame image data into digital signal data, use a Gaussian filter model with 3×3 convolution kernel to denoise the digital signal data, and use a polynomial distortion correction model to perform distortion correction on the denoised digital signal data to obtain the corrected gray value matrix.

[0009] S3: Convert the grayscale matrix into a flame radiation energy signal, use the ART iterative model to iteratively optimize the radiation energy signal, substitute the radiation energy signal after reaching the iteration termination condition into the improved dual-color temperature measurement model to calculate the temperature of each pixel, remove outliers and supplement with the average temperature of adjacent valid pixels to obtain the corrected pixel temperature matrix.

[0010] S4: Construct a global temperature distribution cloud map of the hydrogen-oxygen flame based on the corrected pixel temperature matrix, and extract the feature parameters of the temperature distribution of the hydrogen-oxygen flame.

[0011] S5: Output the temperature distribution cloud map of the oxyhydrogen flame and the characteristic parameters of the temperature distribution of the oxyhydrogen flame in real time, and simultaneously store the temperature data and time to form a temperature change curve.

[0012] S6: Set multiple calibration points within the normal combustion temperature range of the oxyhydrogen flame, collect the standard temperature corresponding to each calibration point, calculate the relative error between the temperature measured by the observation system and the standard temperature, and verify and calibrate the observed temperature.

[0013] Furthermore, in S2, the specific formula for converting the hydrogen-oxygen flame image data into digital signal data is as follows:

[0014]

[0015] in, For pixels The corresponding digital signal data, This represents the light intensity signal of the hydrogen-oxygen flame image at that pixel. The photoelectric conversion coefficient, is the base compensation coefficient.

[0016] Furthermore, in S2, the specific formula for the Gaussian filtering model with the 3×3 convolution kernel is as follows:

[0017]

[0018] in: The Gaussian convolution kernel in coordinates The value at that location, denoted as Gaussian standard deviation, ranging from 0.8 to 1.2, and x and y as the horizontal and vertical coordinates of the convolution kernel, respectively, ranging from [-1, 1].

[0019] Furthermore, in S2, the specific formula for the polynomial distortion correction model is:

[0020]

[0021] in: , These are the x and y coordinates of the distorted pixel in the image coordinate system; , The x and y coordinates of the distort-free pixels in the image coordinate system after correction; , The coordinates of the principal point in the oxyhydrogen flame image; The order of the polynomial; Radial distortion coefficient; The highest order of a polynomial.

[0022] Furthermore, in S3, the specific formula for the ART iterative model is:

[0023]

[0024] in: For the first In the next iteration, the pixel The corresponding optimized value of radiation energy; For the first The optimized radiant energy value corresponding to pixel i in the next iteration; It is a relaxation factor; For pixels The measured radiant energy signal; This is the projection matrix, used to characterize the radiative energy transfer relationship between pixel i and pixel j. This represents the total number of pixels in the oxyhydrogen flame image.

[0025] Furthermore, in S3, the specific formula for converting the grayscale matrix into a flame radiation energy signal is as follows:

[0026]

[0027] in: This represents the flame radiation energy signal; i=1, 2, corresponding to two wavebands respectively; The gray value of band i is converted to the radiant energy conversion coefficient. This is a correction constant for band i. and Obtained through calibration using a halogen tungsten lamp.

[0028] Furthermore, in S3, the specific formula for the improved two-color temperature measurement model is as follows:

[0029]

[0030] in: For pixels The temperature of the hydrogen-oxygen flame at that location; It is Planck's second constant; , These are the wavelengths of the two detection bands, respectively; , The optimized radiant energy signals for the two bands; It is the ratio of the inductance coefficient to the transmittance of the filter under dual-band conditions; To improve the calculation constants of the two-color method, a method for calculating hydrogen-oxygen flames in flame fusion is proposed.

[0031] Furthermore, in S4, the characteristic parameters of the oxyhydrogen flame temperature distribution include: global average temperature, maximum temperature, minimum temperature, and temperature standard deviation. The specific formula for calculating these characteristic parameters is as follows:

[0032]

[0033] in: The average temperature across the entire region; This is the highest temperature; The lowest temperature; denoted as temperature standard deviation; M and N represent the horizontal and vertical pixel counts of the oxyhydrogen flame image, respectively. This is the corrected pixel temperature.

[0034] Furthermore, in S6, the formula for calculating the relative error is:

[0035]

[0036] in: This is a relative error. The temperature measured by the observation system, This corresponds to the standard temperature.

[0037] A system for observing the temperature distribution of an oxyhydrogen flame based on single crystal growth using the flame melting method is characterized in that the system is used to execute any one of the above-described methods for observing the temperature distribution of an oxyhydrogen flame based on single crystal growth using the flame melting method.

[0038] Beneficial effects:

[0039] This invention employs a dual-band high-speed imaging device to acquire full-domain images of an oxyhydrogen flame, combined with pixel-level temperature calculation, enabling visualized observation of the full-domain temperature distribution of the oxyhydrogen flame. This overcomes the limitation of traditional contact temperature measurement, which can only detect local point temperatures, and fully reflects the distribution characteristics of the flame temperature field. Image noise reduction is achieved through a 3×3 Gaussian filter model, and image distortion correction is completed through a polynomial distortion correction model, effectively eliminating image errors caused by equipment and the environment. An ART iterative model is used to optimize the radiant energy signal, combined with a dual-color temperature measurement model improved for the flame fusion method of oxyhydrogen flames, correcting interference from factors such as flame emissivity and equipment induction coefficient. Furthermore, outlier removal and supplementation further improve the accuracy of temperature calculation. This invention utilizes a dual-band high-speed imaging device and multiple algorithm models to achieve rapid processing of image acquisition, signal processing, and temperature calculation. It can output temperature distribution cloud maps and characteristic parameters in real time and simultaneously store temperature change curves, adapting to the real-time observation needs of dynamic combustion in oxyhydrogen flames and providing data support for the dynamic control of single crystal growth processes. The system calibration was completed by calculating the relative error between the observed temperature and the standard temperature, thus establishing a sound temperature measurement verification mechanism, effectively reducing system errors and ensuring the reliability and consistency of temperature observation data. Attached Figure Description

[0040] Figure 1 This is a flowchart of the method of the present invention;

[0041] Figure 2 This is a system flowchart of the present invention;

[0042] Figure 3 This is a comparison chart of the relative errors of various calibration temperature points under different temperature measurement methods for this invention;

[0043] Figure 4 This is a bar chart comparing the defect rates of single crystals grown using different temperature measurement methods according to the present invention.

[0044] Figure 5 This is a measured cloud map of the global temperature distribution of the hydrogen-oxygen flame of the present invention; Detailed Implementation

[0045] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0046] In the description of this invention, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0047] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0048] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0049] Combination Figure 1 As shown, this invention provides a method for observing the temperature distribution of an oxyhydrogen flame based on single-crystal growth using the flame melting method, comprising the following steps:

[0050] S1: Install the dual-band high-speed imaging device at the observation port of the flame fusion single crystal growth device. The dual-band high-speed imaging device can be an ICCD camera or other imaging equipment. The dual-band high-speed imaging device is equipped with two narrow-band filters for the two bands. The lens is aimed at the oxyhydrogen flame area to collect oxyhydrogen flame image data in two bands. The two narrow-band filters are used to filter ambient light, light scattered by raw material dust, and interference light from other impurities in the flame, allowing only the flame radiation light of the target band to pass through and enter the dual-band high-speed imaging device.

[0051] S2: Convert the hydrogen-oxygen flame image data into digital signal data, use a Gaussian filter model with 3×3 convolution kernel to denoise the digital signal data, and use a polynomial distortion correction model to perform distortion correction on the denoised digital signal data to obtain the corrected gray value matrix.

[0052] The specific formula for converting hydrogen-oxygen flame image data into digital signal data is as follows:

[0053]

[0054] in, For pixels The corresponding digital signal data is the basic input for subsequent image processing, converting continuous light intensity into discrete digital quantities to adapt to computer processing. The light intensity signal of the hydrogen-oxygen flame image at this pixel point is directly acquired from a dual-band camera, preserving the true radiation information of the flame without pre-filtering distortion. The photoelectric conversion coefficient linearly maps light intensity into a digital quantity, corrects the camera's photoelectric response deviation, and features linear conversion and fast calculation, ensuring a strict correspondence between grayscale and light intensity. As a basis compensation coefficient, it compensates for system basis noise such as dark current and ambient stray light, eliminates zero drift, and makes measurements more accurate in low temperature or low light areas;

[0055] The specific formula for the Gaussian filtering model with a 3×3 convolution kernel is as follows:

[0056]

[0057] in: The Gaussian convolution kernel in coordinates The values ​​at that point are used for neighborhood weighted average noise reduction, with smooth weight transitions and no hard edge artifacts. The standard deviation is Gaussian, which controls the filtering smoothing intensity. The value ranges from 0.8 to 1.2, which has a narrow range and moderate noise reduction, both removing noise and preserving the details of the flame edge; x and y are the horizontal and vertical coordinates of the convolution kernel, respectively, with a value range of [-1,1]. The small convolution kernel is calculated very quickly and is suitable for real-time processing of high-speed imaging.

[0058] The specific formula for the polynomial distortion correction model is as follows:

[0059]

[0060] in: , The x and y coordinates of the distorted pixels in the image coordinate system are given directly as the original distorted image; no preprocessing is required. , To ensure accurate temperature positioning, the horizontal and vertical coordinates of the distort-free pixels in the image coordinate system are used for precise temperature positioning. The pixel positions correspond one-to-one to ensure the spatial accuracy of temperature measurement. , The coordinates of the principal point of the oxyhydrogen flame image are generally taken from the image center to simplify calibration and make it suitable for rapid installation in industrial settings. The order of the polynomial controls the correction accuracy; a third order is sufficient to correct most of the radial distortion of the lens, balancing accuracy and speed. The radial distortion coefficient is obtained through calibration. It only corrects radial distortion, has low computational cost, and high stability. The highest order of a polynomial.

[0061] S3: Convert the grayscale matrix into a flame radiation energy signal, use the ART iterative model to iteratively optimize the radiation energy signal, substitute the radiation energy signal after reaching the iteration termination condition into the improved dual-color temperature measurement model to calculate the temperature of each pixel, remove outliers and supplement with the average temperature of adjacent valid pixels to obtain the corrected pixel temperature matrix.

[0062] The specific formula for the ART iterative model is as follows:

[0063]

[0064] in: For the first In the next iteration, the pixel The corresponding optimized value of radiation energy; For the first In each iteration, the optimized radiant energy value corresponding to pixel i gradually approaches the true radiant energy, reducing scattering and optical path interference. It serves as a relaxation factor to control the iteration step size, prevent oscillations, has adjustable parameters, and exhibits fast convergence and good stability. For pixels The measured radiation energy signal is based on actual measurements to ensure physical authenticity. The projection matrix is ​​used to characterize the radiative energy transfer relationship between pixel i and pixel j, compensate for flame self-absorption and dust scattering, and significantly improve the accuracy in the high-temperature zone. The iteration range is limited by the total number of pixels in the hydrogen-oxygen flame image to ensure consistent optimization across the entire domain and synchronous optimization across the entire domain, resulting in a more continuous temperature field.

[0065] The specific formula for converting a grayscale matrix into a flame radiation energy signal is as follows:

[0066]

[0067] in: The flame radiation energy signal is the core input for temperature measurement. It converts the image grayscale into physical radiation, which conforms to Planck's law. i=1 and 2 correspond to two bands respectively, realizing dual-color temperature measurement, canceling the influence of emissivity. The dual-band anti-interference is strong and the accuracy is higher than that of the single-color method. The gray value of band i is converted to the radiant energy conversion coefficient. This is a correction constant for band i. and It is obtained through halogen tungsten lamp calibration, can be calibrated on-site, and is compatible with different cameras and lenses, making it highly versatile.

[0068] The specific formula for the improved two-color temperature measurement model is as follows:

[0069]

[0070] in: For pixels The temperature of the hydrogen-oxygen flame at the specified location is measured, and the final temperature value is output. Pixel-level temperature measurement is achieved to realize the global temperature field. It is Planck's second constant, with a value of 1.4388 × 10⁻² m·K, an international standard, ensuring traceability of temperature measurements; , The wavelengths for the two detection bands are selected respectively. The bands with strong flame radiation and low interference, such as 780nm or 880nm, are selected to avoid the absorption peaks of dust and water vapor, resulting in a high signal-to-noise ratio. , The optimized radiant energy signals for the two bands are obtained through ART iterative optimization, which reduces scattering interference and makes the signals purer and more stable for temperature measurement than the original signals. The ratio of the induction coefficient to the transmittance of the filter under dual-band conditions is used to correct the difference in band response between the camera and the filter. To improve the calculation constants of the two-color method, α=1.39 was specifically modified to suit the emissivity characteristics of the oxyhydrogen flame in the flame fusion method, resulting in a smaller error than the general two-color method.

[0071] S4: Construct a global temperature distribution cloud map of the oxyhydrogen flame based on the corrected pixel temperature matrix, and extract the feature parameters of the oxyhydrogen flame temperature distribution.

[0072] The characteristic parameters of the temperature distribution in an oxyhydrogen flame include: global average temperature, maximum temperature, minimum temperature, and temperature standard deviation. The specific formulas for calculating these characteristic parameters are as follows:

[0073]

[0074] in: The average temperature across the entire area is used to assess the overall thermal power of the flame, providing a direct reflection of its heating capacity. This is the highest temperature; The lowest temperature is used to define the high-temperature zone of the flame core and the low-temperature zone of the edge, guiding the matching of the single crystal growth temperature zone. The temperature standard deviation quantifies temperature uniformity; the smaller the value, the more uniform the temperature, directly guiding process stability and improving crystal quality. M and N are the horizontal and vertical pixel counts of the oxyhydrogen flame image, respectively, representing full flame coverage rather than partial coverage. To correct the pixel temperature, avoid raising or lowering feature values ​​for outliers.

[0075] S5: Real-time output of temperature distribution cloud map of oxyhydrogen flame and characteristic parameters of temperature distribution of oxyhydrogen flame, and simultaneous storage of temperature data and time to form temperature change curve;

[0076] S6: Set up multiple calibration points within the normal combustion temperature range of the oxyhydrogen flame, collect the standard temperature corresponding to each calibration point, calculate the relative error between the temperature measured by the observation system and the standard temperature, and verify and calibrate the observed temperature.

[0077] The formula for calculating relative error is:

[0078]

[0079] in: Assuming relative error, evaluate the system's temperature measurement accuracy; The temperature measured by the observation system is directly fed back as the quantity to be calibrated; This corresponds to the standard temperature.

[0080] Combination Figure 2 As shown, the present invention also provides a hydrogen-oxygen flame temperature distribution observation system based on flame melting single crystal growth. The system is used to perform any of the above-mentioned hydrogen-oxygen flame temperature distribution observation methods based on flame melting single crystal growth. The system includes a dual-band high-speed imaging module, a signal processing module, a radiation energy optimization and temperature calculation module, a temperature field construction and feature extraction module, a data output and storage module, and a temperature verification and calibration module.

[0081] Dual-band high-speed imaging module: used to acquire image data of hydrogen-oxygen flame in two bands, providing raw data for subsequent temperature calculation;

[0082] Signal processing module: used to convert image data into digital signal data, and to perform noise reduction and distortion correction to obtain a corrected grayscale matrix;

[0083] Radiation Energy Optimization and Temperature Calculation Module: This module converts grayscale matrix into radiation energy signals and performs pixel temperature calculation and correction through iterative optimization and improvement of the dual-color temperature measurement model.

[0084] Temperature field construction and feature extraction module: used to construct a global temperature distribution cloud map of an oxyhydrogen flame and extract the feature parameters of the temperature distribution;

[0085] Data output and storage module: used to output temperature distribution cloud map and characteristic parameters in real time, and store temperature data and time information to form temperature change curve;

[0086] Temperature verification and calibration module: used to set calibration points and calculate relative errors, complete the verification and calibration of observed temperatures, and improve temperature measurement accuracy.

[0087] Example 1

[0088] according to Figures 3-5 As shown in the figure, this embodiment is applied to the flame fusion growth process of titanium dioxide single crystals. The growth of titanium dioxide single crystals requires high uniformity of the temperature field of the oxyhydrogen flame, and the normal combustion temperature range of the oxyhydrogen flame is 1800℃~2200℃. The specific implementation steps are as follows:

[0089] S1: Install the dual-band high-speed imaging device at the observation port of the flame fusion titanium dioxide single crystal growth device to meet the flame observation requirements of titanium dioxide single crystal growth. Set the detection wavelengths λ1=780nm and λ2=880nm corresponding to the two narrowband filters. Precisely align the device lens with the oxyhydrogen flame combustion zone, adjust the device focal length and exposure parameters to make the flame edge and core area images clearly distinguishable, and collect high-speed image data of the oxyhydrogen flame under dual-band conditions. Set the frame rate to 120fps to adapt to the dynamic change characteristics of flame temperature during titanium dioxide growth.

[0090] S2: According to the formula

[0091]

[0092] The acquired image light intensity signal is converted into digital signal data. Combined with the calibration parameters of the titanium dioxide-grown imaging equipment, the photoelectric conversion coefficient k=0.95 and the substrate compensation coefficient b=0.03 are set. A Gaussian filter model with a 3×3 convolution kernel is used for noise reduction. For the noise characteristics of titanium dioxide flame imaging, the Gaussian standard deviation σ=0.9 is set, and x and y are taken as [-1,0,1]. Then, a polynomial distortion correction model is used for distortion removal. The principal point coordinates (x0,y0) are set as the image center pixel coordinates (1024,768), the highest order of the polynomial N=3, and the radial distortion coefficients k1=0.0009, k2=0.0004, and k3=0.0002 to eliminate image distortion caused by the equipment lens and installation. The corrected gray value matrix is ​​obtained after processing.

[0093] S3: Through high-temperature calibration experiments with halogen tungsten lamps, the grayscale values-radiation energy conversion coefficients k1=1.15 and k2=1.08 for the radiation characteristics of titanium dioxide flames adapted to two bands were obtained, with correction constants b1=0.06 and b2=0.05. According to the formula...

[0094]

[0095] The grayscale matrix is ​​converted into a flame radiation energy signal; the radiation energy signal is iteratively optimized using an ART iterative model, and a relaxation factor is set. =0.75, the iteration termination condition is that the difference between two adjacent iterations of radiation energy signals is less than 0.009, improving the accuracy of radiation energy signals; the optimized radiation energy signals E1'(x,y) and E2'(x,y) are substituted into the improved two-color temperature measurement model, and η=1.03 and α=1.39 are fixed, where the special calculation constant for the oxyhydrogen flame of the flame melting method, C2=1.4388×10⁻²m・K, are used to calculate the temperature of each pixel; the 3σ criterion is used to remove temperature outliers and jump values ​​that are prone to occur at the flame edge during titanium dioxide growth, and the outlier positions are supplemented with the average temperature of effective pixels within the adjacent 3×3 range, to obtain the corrected pixel temperature matrix.

[0096] S4: Based on the corrected pixel temperature matrix, Matlab software was used to draw a cloud map of the global temperature distribution of the oxyhydrogen flame. According to the process analysis requirements of titanium dioxide single crystal growth, the core characteristic parameters of temperature distribution were extracted: global average temperature, maximum temperature, minimum temperature and temperature standard deviation. Each parameter was calculated strictly according to the corresponding formula to provide quantitative indicators for process control.

[0097] S5: The industrial touch screen outputs real-time cloud maps of the temperature distribution of the oxyhydrogen flame and extracted feature parameters. The screen is adapted to the visualization needs of the production site and supports magnified viewing of local temperatures. At the same time, an industrial-grade data storage module is used to store the pixel temperature data and timestamps at each moment. The storage frequency is consistent with the imaging frame rate. Based on the stored data, the temperature change curve of the oxyhydrogen flame is automatically drawn, realizing the dynamic monitoring of the temperature field in the entire process of titanium dioxide single crystal growth, which makes it easy for process personnel to trace the temperature change pattern.

[0098] S6: Within the normal combustion temperature range of 1800℃~2200℃ for titanium dioxide single crystal growth in an oxyhydrogen flame, six calibration points were uniformly set at temperatures of 1800℃, 1900℃, 2000℃, 2050℃, 2100℃, and 2200℃. The 2000℃~2050℃ range is the critical melting temperature range for titanium dioxide powder, and the calibration points were set more densely. A standard high-temperature infrared thermometer (accuracy ±0.5%) was used to collect the standard temperature at each calibration point, and the observed temperature of this observation system was also collected. The temperature was then calculated according to the formula...

[0099]

[0100] The relative error of each calibration point is calculated. In this embodiment, the relative error of each calibration point is less than 1.8%, and the relative error of the calibration points in the critical melting temperature range is controlled within 1.2%, which far meets the temperature measurement accuracy requirements for titanium dioxide single crystal growth. If the relative error of a certain calibration point exceeds the preset threshold (2%), the gray-scale-radiation energy conversion coefficient and photoelectric conversion coefficient of the system are finely adjusted to complete the system calibration and ensure the reliability of the temperature measurement data.

[0101] Example 2

[0102] This embodiment applies to the flame fusion growth process of sapphire single crystals. Sapphire single crystal growth requires high uniformity of the temperature field in the oxyhydrogen flame, and the normal combustion temperature range of the oxyhydrogen flame is 2000℃~2500℃. The specific implementation steps are as follows:

[0103] S1: Install the dual-band high-speed imaging device at the observation port of the flame fusion sapphire single crystal growth device to meet the flame observation requirements of sapphire single crystal growth. Set the detection wavelengths λ1=800nm ​​and λ2=900nm corresponding to the two narrowband filters. Precisely align the device lens with the oxyhydrogen flame combustion zone, adjust the device focal length and exposure parameters to make the flame edge and core area images clearly distinguishable, and acquire high-speed image data of the oxyhydrogen flame under dual-band conditions. Set the frame rate to 100fps to adapt to the dynamic changes in flame temperature during titanium dioxide growth.

[0104] S2: According to the formula

[0105]

[0106] The acquired image light intensity signal is converted into digital signal data. Combined with the calibration parameters of the imaging equipment grown from titanium dioxide, the photoelectric conversion coefficient k=0.98 and the substrate compensation coefficient b=0.02 are set. A Gaussian filter model with a 3×3 convolution kernel is used for noise reduction. For the noise characteristics of sapphire single crystal flame imaging, the Gaussian standard deviation σ=1.0 is set, and x and y are taken as [-1,0,1]. Then, a polynomial distortion correction model is used for distortion removal. The principal point coordinates (x0,y0) are set as the image center pixel coordinates (1024,768), the highest order of the polynomial N=3, and the radial distortion coefficients k1=0.001, k2=0.0005, and k3=0.0001 to eliminate the image distortion caused by the equipment lens and installation. The corrected gray value matrix is ​​obtained after processing.

[0107] S3: Through high-temperature calibration experiments with halogen tungsten lamps, the grayscale values-radiation energy conversion coefficients k1=1.2 and k2=1.1 for two bands adapted to the radiation characteristics of titanium dioxide flames were obtained, along with correction constants b1=0.05 and b2=0.04. According to the formula...

[0108]

[0109] The grayscale matrix was converted into a flame radiation energy signal. The radiation energy signal was iteratively optimized using the ART iterative model, with a relaxation factor ε_ART=0.8 and the iteration termination condition being that the difference between two adjacent iterations of the radiation energy signal is less than 0.01, thus improving the accuracy of the radiation energy signal. The optimized radiation energy signals E1'(x,y) and E2'(x,y) were substituted into the improved dual-color temperature measurement model, with η=1.05 and α=1.39 fixed. The specific calculation constants for the oxyhydrogen flame in the flame melting method, C2=1.4388×10⁻²m・K, were used to calculate the temperature of each pixel. The 3σ criterion was used to remove temperature outliers and jump values ​​that are prone to occur at the flame edge during titanium dioxide growth. The outlier positions were supplemented with the average temperature of effective pixels within a 3×3 range, resulting in the corrected pixel temperature matrix.

[0110] S4: Based on the corrected pixel temperature matrix, Matlab software is used to draw the global temperature distribution cloud map of the oxyhydrogen flame. The temperature distribution feature parameters are extracted by feature extraction algorithm, including global average temperature, maximum temperature, minimum temperature and temperature standard deviation. Each parameter is calculated according to the corresponding formula.

[0111] S5: The industrial touch screen outputs real-time cloud maps of the temperature distribution of the oxyhydrogen flame and extracted feature parameters. The screen is adapted to the visualization needs of the production site and supports magnified viewing of local temperatures. At the same time, an industrial-grade data storage module is used to synchronously store the pixel temperature data and timestamps at each moment. The storage frequency is consistent with the imaging frame rate. Based on the stored data, the temperature change curve of the oxyhydrogen flame is automatically drawn, realizing the dynamic monitoring of the temperature field in the sapphire single crystal growth process, which makes it easy for process personnel to trace the temperature change pattern.

[0112] S6: Within the normal combustion temperature range of an oxyhydrogen flame (1800℃~2200℃), five calibration points are evenly set up with temperatures of 2000℃, 2100℃, 2200℃, 2300℃, 2400℃, and 2500℃ respectively. A standard high-temperature thermometer is used to collect the standard temperature at each calibration point, and the observed temperature of this observation system is also collected. The temperature is then calculated according to the formula...

[0113]

[0114] The relative error of each calibration point is calculated. In this embodiment, the relative error of each calibration point is less than 2%, and the relative error of the calibration points in the critical melting temperature range is controlled within 1.2%, which far meets the temperature measurement accuracy requirements for sapphire single crystal growth. If the relative error of a certain calibration point exceeds the preset threshold (2%), the gray-scale-radiation energy conversion coefficient and photoelectric conversion coefficient of the system are finely adjusted to complete the system calibration and ensure the reliability of the temperature measurement data.

[0115] In the description of this specification, specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.

[0116] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for observing the temperature distribution of an oxyhydrogen flame based on single-crystal growth using the flame melting method, characterized in that, Includes the following steps: S1: Install the dual-band high-speed imaging device at the observation hole of the flame melting single crystal growth device. The dual-band high-speed imaging device is equipped with two narrowband filters in the dual-band. The lens is aimed at the oxyhydrogen flame area to collect oxyhydrogen flame image data in the two bands. S2: Convert the hydrogen-oxygen flame image data into digital signal data, use a Gaussian filter model with 3×3 convolution kernel to denoise the digital signal data, and use a polynomial distortion correction model to perform distortion correction on the denoised digital signal data to obtain the corrected gray value matrix. S3: Convert the grayscale matrix into a flame radiation energy signal, use the ART iterative model to iteratively optimize the radiation energy signal, substitute the radiation energy signal after reaching the iteration termination condition into the improved dual-color temperature measurement model to calculate the temperature of each pixel, remove outliers and supplement with the average temperature of adjacent valid pixels to obtain the corrected pixel temperature matrix. S4: Construct a global temperature distribution cloud map of the hydrogen-oxygen flame based on the corrected pixel temperature matrix, and extract the feature parameters of the hydrogen-oxygen flame temperature distribution. S5: Output the temperature distribution cloud map of the oxyhydrogen flame and the characteristic parameters of the temperature distribution of the oxyhydrogen flame in real time, and simultaneously store the temperature data and time to form a temperature change curve. S6: Set multiple calibration points within the normal combustion temperature range of the oxyhydrogen flame, collect the standard temperature corresponding to each calibration point, calculate the relative error between the temperature measured by the observation system and the standard temperature, and verify and calibrate the observed temperature.

2. The method for observing the temperature distribution of an oxyhydrogen flame based on single crystal growth using the flame melting method according to claim 1, characterized in that, In S2, the specific formula for converting the hydrogen-oxygen flame image data into digital signal data is as follows: in, For pixels The corresponding digital signal data; This represents the light intensity signal of the hydrogen-oxygen flame image at that pixel. The photoelectric conversion coefficient; is the base compensation coefficient.

3. The method for observing the temperature distribution of an oxyhydrogen flame based on single crystal growth using the flame melting method according to claim 1, characterized in that, In S2, the specific formula for the Gaussian filtering model with the 3×3 convolution kernel is: in: The Gaussian convolution kernel in coordinates The value at that location; is the Gaussian standard deviation, with a value of 0.8 to 1.2; x and y are the horizontal and vertical coordinates of the convolution kernel, respectively, with a value range of [-1, 1].

4. The method for observing the temperature distribution of an oxyhydrogen flame based on flame melting single crystal growth according to claim 1, characterized in that, In S2, the specific formula for the polynomial distortion correction model is: in: , These are the x and y coordinates of the distorted pixel in the image coordinate system; , The x and y coordinates of the distort-free pixels in the image coordinate system after correction; , The coordinates of the principal point in the oxyhydrogen flame image; The order of the polynomial; Radial distortion coefficient; The highest order of a polynomial.

5. The method for observing the temperature distribution of an oxyhydrogen flame based on single crystal growth using the flame melting method according to claim 1, characterized in that, In S3, the specific formula for the ART iterative model is: in: For the first In the next iteration, the pixel The corresponding optimized value of radiation energy; For the first The optimized radiant energy value corresponding to pixel i in the next iteration; It is a relaxation factor; For pixels The measured radiant energy signal; This is the projection matrix, used to characterize the radiative energy transfer relationship between pixel i and pixel j. This represents the total number of pixels in the oxyhydrogen flame image.

6. The method for observing the temperature distribution of an oxyhydrogen flame based on single crystal growth using the flame melting method according to claim 1, characterized in that, In S3, the specific formula for converting the grayscale matrix into a flame radiation energy signal is as follows: in: This represents the flame radiation energy signal; i=1, 2, corresponding to two wavebands respectively; The gray value of band i is converted to the radiant energy conversion coefficient. This is a correction constant for band i. and Obtained through calibration using a halogen tungsten lamp.

7. The method for observing the temperature distribution of an oxyhydrogen flame based on flame melting single crystal growth according to claim 1, characterized in that, In S3, the specific formula for the improved two-color temperature measurement model is: in: For pixels The temperature of the hydrogen-oxygen flame at that location; It is Planck's second constant; , These are the wavelengths of the two detection bands, respectively; , The optimized radiant energy signals for the two bands; It is the ratio of the inductance coefficient to the transmittance of the filter under dual-band conditions; To improve the calculation constants of the two-color method, a method for calculating hydrogen-oxygen flames in flame fusion is proposed.

8. The method for observing the temperature distribution of an oxyhydrogen flame based on flame melting single crystal growth according to claim 1, characterized in that, In S4, the characteristic parameters of the oxyhydrogen flame temperature distribution include: global average temperature, maximum temperature, minimum temperature, and temperature standard deviation. The specific formula for calculating these characteristic parameters is as follows: in: The average temperature across the entire region; This is the highest temperature; The lowest temperature; denoted as temperature standard deviation; M and N represent the horizontal and vertical pixel counts of the oxyhydrogen flame image, respectively. This is the corrected pixel temperature.

9. The method for observing the temperature distribution of an oxyhydrogen flame based on single crystal growth using the flame melting method according to claim 1, characterized in that, In S6, the formula for calculating the relative error is: in: This is relative error; The temperature measured by the observation system; This corresponds to the standard temperature.

10. A system for observing the temperature distribution of an oxyhydrogen flame based on single-crystal growth using the flame melting method, characterized in that, The system is used to perform the method for observing the temperature distribution of an oxyhydrogen flame based on single crystal growth by flame melting, as described in any one of claims 1-9.

Images