A method and system for surface modification based silicon carbide crystal diameter expansion

Abstract

The application relates to the technical field of image processing, and relates to a silicon carbide crystal diameter expansion method and system based on surface modification, which comprises the following steps: using an X-ray device to monitor image data of a to-be-monitored silicon carbide crystal to obtain an original ray image; performing noise self-adaptive filtering on the original ray image to obtain a target ray image; identifying a current crystal diameter expansion state according to the target ray image to obtain original crystal diameter expansion data; performing diameter expansion speed fitting based on the original crystal diameter expansion data to obtain a current diameter expansion speed; and adjusting the diameter expansion temperature of a crystal diameter expansion device according to the current diameter expansion speed and an optimal diameter expansion speed to obtain an adjusted diameter expansion device. The application can improve the crystal quality and growth efficiency in the diameter expansion process and reduce process fluctuations caused by speed mismatch.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a method and system for expanding the diameter of silicon carbide crystals based on surface modification. Background Technology

[0002] As a third-generation wide-bandgap semiconductor material, silicon carbide crystal has irreplaceable application value in fields such as power electronics and high-frequency devices. The preparation of large-size, high-quality silicon carbide substrates depends on the radial diameter expansion technology in the crystal growth process. Surface modification schemes can effectively promote uniform radial diameter expansion of the crystal. However, the optimization of the diameter expansion rate, real-time monitoring and precise control of the growth process directly determine the defect density and growth efficiency of the crystal, which are key technologies for realizing the industrialization of high-quality, large-size silicon carbide crystals.

[0003] Traditional techniques typically employ a fixed, empirically-based diameter expansion rate combined with manual visual observation of the growth interface, followed by offline crystal quality testing after growth. The drawbacks of this approach are that a fixed diameter expansion rate cannot be dynamically adjusted according to different modification schemes, which can easily lead to excessively rapid diameter expansion causing edge polycrystalline growth and an increase in inclusions, or excessively slow expansion reducing growth efficiency. Furthermore, relying on manual observation makes it impossible to accurately measure the crystal radius in real time, and traditional image filtering methods can blur the crystal edges, causing cumulative radius measurement errors. Ultimately, it is difficult to achieve closed-loop feedback control of the diameter expansion rate. Summary of the Invention

[0004] This invention provides a method for expanding the diameter of silicon carbide crystals based on surface modification and a computer-readable storage medium. Its main purpose is to improve the crystal quality and growth efficiency during the diameter expansion process and reduce process fluctuations caused by speed mismatch.

[0005] To achieve the above objectives, the present invention provides a method for increasing the diameter of silicon carbide crystals based on surface modification, comprising:

[0006] The diameter expansion rate of the pre-obtained original modification scheme is optimized to obtain the optimal diameter expansion rate;

[0007] The crystal to be expanded is subjected to crystal expansion according to the original modification scheme and the pre-constructed crystal expansion equipment to obtain the silicon carbide crystal to be monitored. The crystal expansion equipment includes an X-ray device.

[0008] X-ray equipment was used to monitor the image data of the silicon carbide crystal under monitoring, and the original X-ray image set was obtained.

[0009] The original ray images are extracted sequentially from the original ray image set, and the extracted original ray images are subjected to noise adaptive filtering to obtain the target ray image;

[0010] The current crystal diameter expansion status is identified based on the target ray image to obtain the original crystal diameter expansion data, which includes: expansion timestamp and original crystal radius;

[0011] The original crystal diameter expansion data corresponding to each original X-ray image are summarized to obtain the original crystal diameter expansion dataset. Based on the original crystal diameter expansion dataset, the diameter expansion rate is fitted to obtain the current diameter expansion rate.

[0012] The crystal expansion equipment is adjusted by regulating the expansion temperature based on the current expansion rate and the optimal expansion rate, thus obtaining an adjustable expansion equipment and completing the silicon carbide crystal expansion.

[0013] Optionally, the step of optimizing the diameter expansion rate of the pre-obtained original modification scheme to obtain the optimal diameter expansion rate includes:

[0014] The preset expansion speed range is discretized to obtain a discrete expansion speed set;

[0015] Discrete diameter expansion velocities are extracted sequentially from the discrete diameter expansion velocity set. Based on the original modification scheme and the extracted discrete diameter expansion velocities, the diameter of the pre-obtained experimental silicon carbide crystal is expanded to obtain the experimental diameter-expanded crystal.

[0016] Defect detection was performed on the experimental expanded-diameter crystal to obtain the experimental defect values;

[0017] The test defect values ​​corresponding to each discrete diameter expansion rate are summarized to obtain the test defect value set;

[0018] The optimal expansion speed is determined within the expansion speed range based on the set of experimental defect values.

[0019] Optionally, the step of performing noise adaptive filtering on the extracted original ray image to obtain the target ray image includes:

[0020] Global gradient data is calculated on the extracted original ray image to obtain the global gradient mean and global noise variance;

[0021] The original ray image is divided into local windows to obtain multiple windowed ray images;

[0022] Noise deconstruction is performed using multiple windowed ray images, global gradient mean, and global noise variance to obtain quantum noise ray images and scattered noise ray images;

[0023] Quantum noise filtering is applied to the quantum noise ray image to obtain a quantum filtered ray image;

[0024] The scattered noise ray image is filtered to obtain the scattered noise filtered ray image;

[0025] The target ray image is obtained by synthesizing the quantum-filtered ray image and the scattered ray image.

[0026] Optionally, the step of using multiple windowed ray images, the global gradient mean, and the global noise variance to perform noise deconstruction to obtain quantum noise ray images and scattered noise ray images includes:

[0027] Perform the following operation on each of the multiple windowed ray images:

[0028] Window gradient data calculation is performed on the windowed ray image to obtain the local gradient mean and local noise variance;

[0029] The noise decision threshold is calculated based on the global gradient mean, global noise variance, local gradient mean, and local noise variance.

[0030] The windowed ray image is labeled according to the noise decision threshold to obtain the labeled ray image;

[0031] Adaptive wavelet decomposition is performed on the marked ray image to obtain multiple original wavelet decomposed images;

[0032] Identify high-frequency and low-frequency wavelet decomposition image groups in multiple original wavelet decomposition images;

[0033] Noise reconstruction is performed on the high-frequency wavelet decomposition image group to obtain a quantum noise ray image, and noise reconstruction is performed on the low-frequency wavelet decomposition image group to obtain a scattered noise ray image.

[0034] Optionally, the step of performing quantum noise filtering on the quantum noise ray image to obtain a quantum filtered ray image includes:

[0035] Poisson deviation is calculated on quantum noise ray images to obtain the global Poisson deviation.

[0036] Multiple sliding quantum noise regions are obtained by sliding the quantum noise ray image through a preset sliding window.

[0037] Sliding quantum noise regions are extracted sequentially from multiple sliding quantum noise regions, and Poisson deviation is calculated for the extracted sliding quantum noise regions to obtain the local Poisson deviation.

[0038] Anscombe transform is performed on the extracted sliding quantum noise region based on the global and local Poisson deviations to obtain the sliding Gaussian noise region.

[0039] By summing up the sliding Gaussian noise regions corresponding to each sliding quantum noise region, multiple sliding Gaussian noise regions are obtained.

[0040] A global transformed ray image is obtained by merging multiple sliding Gaussian noise regions.

[0041] Gaussian filtering is applied to the globally transformed ray image to obtain a globally filtered ray image;

[0042] An inverse Anscombe transform is performed on the global filtered ray image to obtain the quantum filtered ray image.

[0043] Optionally, the step of filtering the scattering noise ray image to obtain a filtered scattering ray image includes:

[0044] Multiple sliding noise regions are obtained by sliding the scattering noise ray image using a sliding window.

[0045] For each of the multiple sliding scattering noise regions, the following operation is performed:

[0046] The set of scattered noise pixels in the sliding scattering noise region is identified, and bilateral filtering is performed using the set of scattered noise pixels to obtain the sliding filtering region.

[0047] By summarizing the sliding filter regions, multiple sliding filter regions are obtained, and a scattering filter ray image is generated based on these multiple sliding filter regions.

[0048] Optionally, the step of performing bilateral filtering using the set of scattered noise pixels to obtain a sliding filtering region includes:

[0049] Scattered noise pixels are extracted sequentially from the set of scattered noise pixels, and the extracted scattered noise pixels are recorded as the center noise pixel.

[0050] Based on the central noise pixel, identify the neighborhood noise pixel set within the scattered noise pixel set;

[0051] The filtering weights of the central noise pixel are calculated based on the neighborhood noise pixel set to obtain the bilateral filtering weight set;

[0052] The weighted gray value is obtained by using the bilateral filter weight set and the neighborhood noise pixel set to perform gray-level weighted calculation.

[0053] The grayscale value of the center noise pixel is updated using the weighted filtered grayscale value to obtain the center filtered pixel;

[0054] The central filtered pixels corresponding to each central noise pixel are summed to obtain the central filtered pixel set;

[0055] The sliding scattering noise region is updated using the central filtered pixel set to obtain the sliding filtered region.

[0056] Optionally, the step of calculating the filtering weights for the central noise pixel based on the neighborhood noise pixel set to obtain a bilateral filtering weight set includes:

[0057] The average gradient of the sliding scattering noise region is calculated based on the set of scattering noise pixels to obtain the average region gradient.

[0058] Obtain the set of scattered gray values ​​corresponding to the set of scattered noise pixels, and use the set of scattered gray values ​​to calculate the gray value variance of the region.

[0059] For each neighboring noisy pixel in the neighborhood noise pixel set, perform the following operation:

[0060] The grayscale difference weight and pixel space weight between the neighboring noise pixels and the center noise pixels are calculated using the average regional gradient and regional grayscale variance.

[0061] The mean of the grayscale difference weight and the pixel space weight is calculated to obtain the bilateral filter weight;

[0062] The bilateral filter weights corresponding to each neighboring noisy pixel are summed to obtain the bilateral filter weight set.

[0063] Optionally, the step of identifying the current crystal diameter expansion state based on the target ray image to obtain the original crystal diameter expansion data includes:

[0064] Edge detection is performed based on the target ray image to obtain the expanded crystal contour, and the coordinates of the contour center in the expanded crystal contour are identified.

[0065] The contour coordinates of the expanded crystal are uniformly extracted to obtain the contour edge coordinate set;

[0066] Based on the coordinates of the contour center, the distance to each contour edge coordinate in the contour edge coordinate set is calculated to obtain the contour edge distance set.

[0067] The original crystal radius is obtained by averaging the distance set of the contour edges;

[0068] Obtain the expansion timestamp corresponding to the original X-ray image, merge the expansion timestamp and the original crystal radius to obtain the original crystal expansion data.

[0069] To achieve the above objectives, the present invention also provides a silicon carbide crystal diameter expansion system based on surface modification, comprising:

[0070] The silicon carbide crystal modification module is used to optimize the expansion rate of the pre-acquired original modification scheme to obtain the optimal expansion rate. Based on the original modification scheme and the pre-constructed crystal expansion equipment, the module expands the diameter of the pre-acquired crystal to be expanded to obtain the silicon carbide crystal to be monitored. The crystal expansion equipment includes an X-ray device.

[0071] The X-ray image acquisition module is used to monitor the image data of the silicon carbide crystal under monitoring using X-ray equipment, obtain the original X-ray image set, extract the original X-ray images sequentially from the original X-ray image set, and perform noise adaptive filtering on the extracted original X-ray images to obtain the target X-ray image.

[0072] The crystal data extraction module is used to identify the current crystal diameter expansion state based on the target ray image and obtain the original crystal diameter expansion data. The original crystal diameter expansion data includes: diameter expansion timestamp and original crystal radius. The original crystal diameter expansion data corresponding to each original ray image is summarized to obtain the original crystal diameter expansion dataset. Based on the original crystal diameter expansion dataset, the diameter expansion rate is fitted to obtain the current diameter expansion rate.

[0073] The diameter expansion speed optimization module is used to adjust the diameter expansion temperature of the crystal diameter expansion equipment according to the current diameter expansion speed and the optimal diameter expansion speed, so as to obtain the adjustable diameter expansion equipment.

[0074] To address the above problems, the present invention also provides an electronic device, the electronic device comprising:

[0075] Memory, storing at least one instruction;

[0076] The processor executes the instructions stored in the memory to implement the surface-modified silicon carbide crystal diameter expansion method described above.

[0077] To address the aforementioned problems, the present invention also provides a computer-readable storage medium storing at least one instruction, which is executed by a processor in an electronic device to implement the aforementioned silicon carbide crystal diameter expansion method based on surface modification.

[0078] To address the problems described in the background section, this invention first optimizes the diameter expansion rate of the original modification scheme to obtain the optimal expansion rate. This step involves optimizing the diameter expansion rate of the original modification scheme, combining defect detection results with curve fitting and weighting within a discretized speed range to ultimately determine the optimal expansion rate. Compared to the traditional method of setting a fixed expansion rate based on experience, this method dynamically balances crystal quality and growth efficiency, avoiding excessively high speeds that lead to increased edge polycrystalline material or inclusions, and also preventing excessively low speeds that reduce production efficiency. Furthermore, this scheme performs adaptive noise filtering on the original X-ray image to obtain the target X-ray image. This step, through adaptive noise filtering of the original X-ray image, deconstructs the image into two parts: quantum noise and scattering noise, and employs targeted filtering strategies for each, overcoming the limitations of traditional methods. Traditional Gaussian or median filtering suffers from limitations such as inability to distinguish noise types and tendency to blur crystal edges. This invention effectively suppresses high-frequency quantum noise and low-frequency Compton scattering noise while preserving the crystal outline to the maximum extent, providing a high-quality image foundation for subsequent accurate crystal radius measurement. Finally, the crystal expansion equipment's expansion temperature is adjusted based on the current and optimal expansion speeds, resulting in an regulated expansion device. This step uses a PID controller to regulate the crystal expansion equipment temperature based on the deviation between the current and optimal expansion speeds, automatically bringing the expansion speed closer to the optimal value. Compared to the traditional method of manual temperature adjustment, this method achieves closed-loop feedback control, enabling real-time correction of deviations during the growth process and ensuring that the expansion process always operates under optimal conditions, improving the stability and consistency of the expansion process. Therefore, this invention can improve crystal quality and growth efficiency during the expansion process and reduce process fluctuations caused by speed mismatch. Attached Figure Description

[0079] Figure 1 This is a schematic flowchart of a silicon carbide crystal diameter expansion method based on surface modification provided in an embodiment of the present invention;

[0080] Figure 2 A functional block diagram of a surface-modified silicon carbide crystal diameter expansion system provided in an embodiment of the present invention;

[0081] Figure 3 This is a schematic diagram of an electronic device that implements the surface-modified silicon carbide crystal diameter expansion method according to an embodiment of the present invention.

[0082] Explanation of reference numerals in the attached figures:

[0083] 10. Electronic device; 11. Processor; 12. Memory; 13. Bus.

[0084] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0085] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0086] This application provides a method for increasing the diameter of silicon carbide crystals based on surface modification. The execution entity of this method includes, but is not limited to, at least one electronic device configured to execute the method provided in this application, such as a server or a terminal. In other words, the method can be executed by software or hardware installed on a terminal device or a server device, and the software may be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0087] Reference Figure 1 The diagram shown is a schematic flowchart of a silicon carbide crystal diameter expansion method based on surface modification according to an embodiment of the present invention. In this embodiment, the silicon carbide crystal diameter expansion method based on surface modification includes:

[0088] S1. Optimize the diameter expansion rate of the pre-obtained original modification scheme to obtain the optimal diameter expansion rate.

[0089] Understandably, the original modification scheme refers to a scheme that chemically modifies the surface of the seed crystal of silicon carbide, such as preparing a coating of a specific material on the surface of the seed crystal to promote uniform radial expansion of the silicon carbide crystal.

[0090] Optionally, a tantalum carbide (TaC) coating can be used as the initial modification scheme. Tantalum carbide possesses a high melting point, excellent chemical inertness, and a thermal expansion coefficient close to that of silicon carbide. This allows it to maintain structural stability for extended periods in the 2100–2300℃ growth environment of subsequent crystal expansion steps (such as crystal expansion using physical vapor transport), thereby improving the uniformity of the expansion process. The optimal expansion rate refers to the expansion rate obtained after optimization, where the expansion rate is the increase in the radius of the silicon carbide crystal per unit time, expressed in millimeters per hour (mm / h). Since different expansion rates affect the dislocation density, the number of microtube defects, and the stability of the expansion interface within the silicon carbide crystal, excessively fast expansion rates can lead to edge polycrystalline structures and increased inclusions, while excessively slow expansion rates prolong the expansion time and reduce efficiency. Therefore, it is necessary to determine the optimal expansion rate under this initial modification scheme to maximize expansion efficiency while meeting the quality requirements of the silicon carbide crystal.

[0091] Specifically, the optimization of the diameter expansion rate of the pre-obtained original modification scheme to obtain the optimal diameter expansion rate includes:

[0092] The preset expansion speed range is discretized to obtain a discrete expansion speed set;

[0093] Discrete diameter expansion velocities are extracted sequentially from the discrete diameter expansion velocity set. Based on the original modification scheme and the extracted discrete diameter expansion velocities, the diameter of the pre-obtained experimental silicon carbide crystal is expanded to obtain the experimental diameter-expanded crystal.

[0094] Defect detection was performed on the experimental expanded-diameter crystal to obtain the experimental defect values;

[0095] The test defect values ​​corresponding to each discrete diameter expansion rate are summarized to obtain the test defect value set;

[0096] The optimal expansion speed is determined within the expansion speed range based on the set of experimental defect values.

[0097] It should be explained that the diameter expansion rate range refers to a range of diameter expansion rates for silicon carbide crystals that is artificially set. This range can be set based on previous silicon carbide growth experience data, for example, 0.5 mm / h to 5.0 mm / h. The discrete diameter expansion rate set refers to a collection of multiple discrete diameter expansion rates. A discrete diameter expansion rate is obtained by discretizing the diameter expansion rate range. Discrete division means uniformly dividing a continuous diameter expansion rate interval into several discrete values ​​at a certain step size (e.g., 0.5 mm / h), each of which is a discrete diameter expansion rate. The experimental diameter-expanded crystal refers to a test silicon carbide crystal after diameter expansion. This test silicon carbide crystal has the same model as the silicon carbide crystal that needs diameter expansion in this scheme (i.e., the subsequent crystal to be expanded). The diameter expansion method of this test silicon carbide crystal is the same as the method used for the subsequent crystal diameter expansion of the crystal to be expanded. The test defect value refers to the numerical value of the surface defects of the test diameter-expanding crystal. The larger the test defect value, the more serious the defects of the test diameter-expanding crystal. The test defect value can be calculated by performing KOH melt etching on the test diameter-expanding crystal and counting the number of dislocation etch pits per unit area under an optical microscope. The more dislocation etch pits there are, the more defects the test diameter-expanding crystal produces. That is, the number of dislocation etch pits is taken as the test defect value.

[0098] Optionally, the specific method for determining the optimal diameter expansion speed within the diameter expansion speed range based on the test defect value set includes: performing curve fitting based on the test defect value set to obtain a test defect curve, wherein the horizontal axis and vertical axis of the test defect curve represent the test defect value and the discrete diameter expansion speed, respectively;

[0099] The test defect curve is normalized by adjusting the ordinate to obtain the normalized defect curve;

[0100] Receive the expansion efficiency weight and crystal quality weight, and use the expansion efficiency weight and crystal quality weight to weight the normalized defect curve on the horizontal and vertical axes to obtain the expansion weighted curve.

[0101] Identify the maximum point in the diameter expansion weighted curve and determine the optimal diameter expansion rate based on the maximum point.

[0102] It should be explained that the experimental defect curve refers to the curve obtained after curve fitting, which represents the relationship between the experimental defect value and the discrete diameter expansion rate. The normalized defect curve refers to the curve where the ordinate values ​​of the experimental defect curve are scaled to the range of 0 to 1. The minimum-maximum normalization method can be used to normalize each ordinate value in the experimental defect curve; the normalized experimental defect curve is the normalized defect curve. The diameter expansion efficiency weight refers to a manually set value representing the importance of the discrete diameter expansion rate, and the crystal quality weight refers to a manually set value representing the importance of the experimental defect value. These diameter expansion efficiency weights and crystal quality weights can be set according to actual needs, and their sum is 1. The aforementioned diameter expansion weighted curve refers to the curve obtained after weighting the horizontal and vertical coordinates. The specific method for weighting the horizontal and vertical coordinates is as follows: For each coordinate point in the normalized defect curve, the following calculation is performed: ,in, This represents the value obtained after weighted calculation using the horizontal and vertical axes. and These represent the crystal quality weight and the diameter expansion efficiency weight, respectively. The ordinate of the coordinate point represents the normalized test defect value. This represents the minimum and maximum value normalization function. The x-coordinate of the coordinate point represents the discrete expansion velocity.

[0103] S2. Based on the original modification scheme and the pre-constructed crystal diameter expansion equipment, the pre-obtained crystal to be expanded is subjected to crystal diameter expansion to obtain the silicon carbide crystal to be monitored. The crystal diameter expansion equipment includes an X-ray device.

[0104] Understandably, the crystal diameter expansion equipment refers to a growth device used to achieve radial enlargement of silicon carbide crystals, such as a single crystal furnace using physical vapor transport (PVT) method, equipped with a graphite crucible, insulation felt, temperature control system, etc., and the sidewall of the crystal diameter expansion equipment also includes a window for X-ray observation. The crystal to be expanded refers to the silicon carbide seed crystal that needs to be expanded. The silicon carbide crystal to be monitored refers to the crystal undergoing diameter expansion. The X-ray equipment refers to a real-time X-ray imaging system, consisting of an X-ray source, a flat panel detector, and an image acquisition and processing unit. This X-ray equipment is used to penetrate the graphite crucible and acquire a real-time projected image of the silicon carbide crystal growth interface inside the graphite crucible, thereby monitoring the crystal radius.

[0105] S3. Use X-ray equipment to monitor the image data of the silicon carbide crystal to be monitored, and obtain the original X-ray image set.

[0106] It is clear that the raw X-ray image set refers to a collection of multiple raw X-ray images, wherein the raw X-ray image refers to an X-ray transmission projection image captured by an X-ray device at a certain moment.

[0107] S4. Extract the original ray images sequentially from the original ray image set, and perform noise adaptive filtering on the extracted original ray images to obtain the target ray image.

[0108] Understandably, the target X-ray image refers to the original X-ray image after adaptive filtering. Since the original X-ray image contains both high-frequency quantum noise (originating from statistical fluctuations in X-ray photons) and low-frequency Compton scattering noise (originating from inelastic scattering of X-rays as they pass through the graphite crucible and crystal interface), the spectral characteristics and spatial distribution of these two types of noise differ significantly. Traditional techniques typically employ Gaussian filtering or median filtering, but this method cannot distinguish between noise types and easily blurs the crystal edges of the silicon carbide crystal under monitoring while suppressing noise, leading to a decrease in the accuracy of subsequent radius measurements. Therefore, this scheme introduces this noise adaptive filtering step. This step deconstructs the original X-ray image into a quantum noise X-ray image and a scattering noise X-ray image, and then adaptively filters the quantum noise X-ray image and the scattering noise X-ray image respectively, thereby effectively suppressing both types of noise while preserving the crystal outline of the silicon carbide crystal under monitoring to the maximum extent.

[0109] Specifically, the step of performing noise adaptive filtering on the extracted original ray image to obtain the target ray image includes:

[0110] Global gradient data is calculated on the extracted original ray image to obtain the global gradient mean and global noise variance;

[0111] The original ray image is divided into local windows to obtain multiple windowed ray images;

[0112] Noise deconstruction is performed using multiple windowed ray images, global gradient mean, and global noise variance to obtain quantum noise ray images and scattered noise ray images;

[0113] Quantum noise filtering is applied to the quantum noise ray image to obtain a quantum filtered ray image;

[0114] The scattered noise ray image is filtered to obtain the scattered noise filtered ray image;

[0115] The target ray image is obtained by synthesizing the quantum-filtered ray image and the scattered ray image.

[0116] It should be explained that the global gradient mean refers to the arithmetic mean of the gradient values ​​of all pixels in the original ray image. This global gradient mean represents the overall texture complexity of the original ray image. The calculation method for the global gradient mean is as follows: the Sobel operator is used to calculate the horizontal and vertical gradients of each pixel in the original ray image. The squares of the horizontal and vertical gradients are summed, and the square root is taken. The resulting value is the gradient magnitude. The average value of the gradient magnitude corresponding to each pixel is calculated, and this average value is the global gradient mean. The global noise variance refers to the variance of the grayscale values ​​corresponding to all pixels in the original ray image. The windowed ray image refers to the non-overlapping rectangular regions in the original ray image obtained after local window partitioning. Local window partitioning refers to dividing the original image into multiple adjacent rectangular regions of a fixed size; each rectangular region is a windowed ray image.

[0117] Furthermore, the quantum noise ray image refers to the original ray image composed of high-frequency particulate quantum noise, the noise in which originates from statistical fluctuations in X-ray photon counting. The scattering noise ray image refers to the original ray image composed of low-frequency hazy artifacts, the noise in which originates from Compton scattering generated when X-rays pass through high-density interfaces such as graphite crucibles. The specific method for image synthesis based on the quantum-filtered ray image and the scattering-filtered ray image is as follows: the gray values ​​of the quantum-filtered ray image and the scattering-filtered ray image are added pixel by pixel, that is, the gray values ​​of pixels at the same position in both images are added together.

[0118] In detail, the step of using multiple windowed ray images, global gradient mean, and global noise variance to perform noise deconstruction to obtain quantum noise ray images and scattered noise ray images includes:

[0119] Perform the following operation on each of the multiple windowed ray images:

[0120] Window gradient data calculation is performed on the windowed ray image to obtain the local gradient mean and local noise variance;

[0121] The noise decision threshold is calculated based on the global gradient mean, global noise variance, local gradient mean, and local noise variance.

[0122] The windowed ray image is labeled according to the noise decision threshold to obtain the labeled ray image;

[0123] Adaptive wavelet decomposition is performed on the marked ray image to obtain multiple original wavelet decomposed images;

[0124] Identify high-frequency and low-frequency wavelet decomposition image groups in multiple original wavelet decomposition images;

[0125] Noise reconstruction is performed on the high-frequency wavelet decomposition image group to obtain a quantum noise ray image, and noise reconstruction is performed on the low-frequency wavelet decomposition image group to obtain a scattered noise ray image.

[0126] It should be explained that the local gradient mean refers to the arithmetic mean of the gradient values ​​of all pixels in the windowed ray image, and the calculation method for this local gradient mean is the same as that for the global gradient mean described above. The local noise variance refers to the variance of the grayscale values ​​corresponding to all pixels in the windowed ray image.

[0127] Optionally, the specific method for marking the window ray image based on the noise decision threshold is as follows:

[0128] If the noise decision threshold is greater than the preset noise discrimination threshold, the window ray image is labeled with the noise type based on the preset scattering noise label to obtain the labeled ray image;

[0129] If the noise decision threshold is not greater than the noise discrimination threshold, the window ray image is labeled with noise type based on the preset quantum noise label to obtain the labeled ray image.

[0130] It should be explained that the noise decision threshold refers to a value used to determine which type of noise dominates in a windowed ray image. The larger the noise decision threshold, the more prominent the local gradient mean is relative to the global gradient mean, and the larger the local noise variance is relative to the global noise variance. In other words, the windowed ray image is more likely to belong to a region dominated by Compton scattering noise. The formula for calculating the noise decision threshold is as follows: ,in, Indicates the noise decision threshold. Represents the local gradient mean. Represents the global gradient mean. Represents the local noise variance. This represents the global noise variance. The noise discrimination threshold is a manually set critical value used to divide the noise decision threshold into two intervals, corresponding to quantum noise dominance and scattering noise dominance, respectively. This noise region threshold can be set based on statistical analysis of a large number of experimental images or empirical values, for example, 1.0.

[0131] Clearly, when the noise decision threshold is greater than the noise region threshold, it indicates that the energy proportion of Compton scattering noise in the windowed X-ray image is higher than that of quantum noise, meaning that the windowed X-ray image is mainly generated by Compton scattering. The scattering noise label refers to a numerical label artificially set to mark scattering noise, for example, a value of 1. The labeled X-ray image refers to the windowed X-ray image after noise type labeling. When the noise decision threshold is not greater than the noise region threshold, it indicates that the noise is mainly generated by statistical fluctuations in X-ray photons (i.e., quantum noise). The quantum noise label refers to a numerical label artificially set to mark quantum noise, for example, a value of 0.

[0132] Furthermore, the adaptive wavelet decomposition refers to using different wavelet decomposition methods for different noise types. Optionally, if the markers in the marked ray image are scattering noise labels, the sym2 wavelet basis algorithm can be used to perform wavelet decomposition on the marked scattering image; if the markers in the marked ray image are quantum noise labels, the db6 wavelet basis algorithm can be used to perform wavelet decomposition on the marked scattering image. The original wavelet decomposed image refers to the sub-images of each frequency band obtained after adaptive wavelet decomposition.

[0133] It should be explained that the high-frequency wavelet decomposition image group refers to a set of one or more high-frequency wavelet decomposition images, wherein a high-frequency wavelet decomposition image refers to a sub-image containing high-frequency detail information. The low-frequency wavelet decomposition image group refers to a set of one or more low-frequency wavelet decomposition images, wherein a low-frequency wavelet decomposition image refers to a sub-image containing low-frequency detail information. The high-frequency and low-frequency wavelet decomposition image groups are identified by distinguishing them according to the wavelet decomposition level and sub-band name. For example, for the db6 wavelet basis algorithm (three-level decomposition), the image corresponding to the LL sub-band of the third level is a low-frequency wavelet decomposition image, and the sub-images corresponding to all other LH, HL, and HH sub-bands are high-frequency wavelet decomposition images. For the sym2 wavelet basis algorithm (one-level decomposition), the sub-image corresponding to the ll sub-band is a low-frequency wavelet decomposition image, and the sub-images corresponding to the lh, hl, and hh sub-bands are high-frequency wavelet decomposition images. The aforementioned noise reconstruction based on high-frequency wavelet decomposition image groups refers to recombining all sub-images belonging to the high-frequency wavelet decomposition image group into a single spatial domain image through inverse wavelet transform. This spatial domain image is the quantum noise ray image. Since quantum noise is mainly distributed in the high-frequency sub-band in the frequency domain, the image obtained after reconstructing these high-frequency wavelet decomposition images is used as the quantum noise ray image. The noise reconstruction based on low-frequency wavelet decomposition image groups is performed in the same way as the noise reconstruction based on high-frequency wavelet decomposition image groups, and will not be repeated here. Since Compton scattering noise is mainly distributed in the low-frequency sub-band in the frequency domain, the image obtained after reconstructing these low-frequency wavelet decomposition images is used as the scattering noise ray image.

[0134] In detail, the process of performing quantum noise filtering on the quantum noise ray image to obtain a quantum filtered ray image includes:

[0135] Poisson deviation is calculated on quantum noise ray images to obtain the global Poisson deviation.

[0136] Multiple sliding quantum noise regions are obtained by sliding the quantum noise ray image through a preset sliding window.

[0137] Sliding quantum noise regions are extracted sequentially from multiple sliding quantum noise regions, and Poisson deviation is calculated for the extracted sliding quantum noise regions to obtain the local Poisson deviation.

[0138] Anscombe transform is performed on the extracted sliding quantum noise region based on the global and local Poisson deviations to obtain the sliding Gaussian noise region.

[0139] By summing up the sliding Gaussian noise regions corresponding to each sliding quantum noise region, multiple sliding Gaussian noise regions are obtained.

[0140] A global transformed ray image is obtained by merging multiple sliding Gaussian noise regions.

[0141] Gaussian filtering is applied to the globally transformed ray image to obtain a globally filtered ray image;

[0142] An inverse Anscombe transform is performed on the global filtered ray image to obtain the quantum filtered ray image.

[0143] It should be explained that the global Poisson deviation refers to a numerical value used to represent the degree of deviation between the gray-level distribution and the Poisson distribution in a quantum noise ray image. The larger the global Poisson deviation, the greater the deviation from the Poisson distribution. The global Poisson deviation is calculated as follows: first, the gray-level mean and gray-level variance of the quantum noise ray image are calculated, and then the global Poisson deviation is: ,in, This indicates taking the absolute value. Represents the variance of gray levels. This represents the grayscale mean. The sliding window refers to a window used to slide across the quantum noise ray image, and the sliding quantum noise region refers to an image region selected after sliding. The method of obtaining and the meaning of the local Poisson deviation are the same as those of the global Poisson deviation described above, and will not be repeated here.

[0144] Furthermore, the sliding Gaussian noise region refers to the sliding quantum noise region after Anscombe transformation. Anscombe transformation is used to approximate noise following a Poisson distribution into noise following a Gaussian distribution, so that it can be processed subsequently using mature methods such as Gaussian filtering. The specific method of Anscombe transformation is as follows: if the local Poisson deviation is less than the global Poisson deviation, it indicates that the noise in the sliding quantum noise region is closer to the ideal Poisson distribution, and the standard Anscombe transformation can be used. The Anscombe transformation can be performed using the following formula: Where d represents the grayscale value of each pixel within the sliding quantum noise region. If the local Poisson deviation is not less than the global Poisson deviation, it indicates that the noise distribution in the sliding quantum noise region deviates from the ideal Poisson distribution. In this case, the generalized Anscombe transform can be used instead of the standard Anscombe transform. The specific calculation formula for the generalized Anscombe transform is existing technology and will not be elaborated here.

[0145] As is clear, the region merging refers to merging multiple sliding Gaussian noise regions into a complete image, i.e., a globally transformed ray image, according to their corresponding positions. The globally filtered ray image refers to the globally transformed ray image after Gaussian filtering; this step is used to further suppress the Gaussian noise remaining after the Anscombe transform. The inverse Anscombe transform refers to restoring the Gaussian-filtered image from the Gaussian distribution space back to the original Poisson distribution space. The specific calculation method of this inverse Anscombe transform is also existing technology and will not be elaborated here.

[0146] Specifically, the step of filtering the scattering noise ray image to obtain a filtered scattering ray image includes:

[0147] Multiple sliding noise regions are obtained by sliding the scattering noise ray image using a sliding window.

[0148] For each of the multiple sliding scattering noise regions, the following operation is performed:

[0149] The set of scattered noise pixels in the sliding scattering noise region is identified, and bilateral filtering is performed using the set of scattered noise pixels to obtain the sliding filtering region.

[0150] By summarizing the sliding filter regions, multiple sliding filter regions are obtained, and a scattering filter ray image is generated based on these multiple sliding filter regions.

[0151] It should be explained that the "sliding scattering noise region" refers to a selected image region in the scattering noise ray image after sliding. The "scattering noise pixel set" refers to a collection of multiple scattering noise pixels, where each scattering noise pixel refers to a pixel within the sliding scattering noise region. The "sliding filter region" refers to the sliding scattering noise region after bilateral filtering. The above-mentioned generation of a scattering filter ray image based on multiple sliding filter regions means: merging multiple sliding filter regions according to their respective positions into a complete image, which is the scattering filter ray image.

[0152] In detail, the step of performing bilateral filtering using the set of scattered noise pixels to obtain the sliding filtering region includes:

[0153] Scattered noise pixels are extracted sequentially from the set of scattered noise pixels, and the extracted scattered noise pixels are recorded as the center noise pixel.

[0154] Based on the central noise pixel, identify the neighborhood noise pixel set within the scattered noise pixel set;

[0155] The filtering weights of the central noise pixel are calculated based on the neighborhood noise pixel set to obtain the bilateral filtering weight set;

[0156] The weighted gray value is obtained by using the bilateral filter weight set and the neighborhood noise pixel set to perform gray-level weighted calculation.

[0157] The grayscale value of the center noise pixel is updated using the weighted filtered grayscale value to obtain the center filtered pixel;

[0158] The central filtered pixels corresponding to each central noise pixel are summed to obtain the central filtered pixel set;

[0159] The sliding scattering noise region is updated using the central filtered pixel set to obtain the sliding filtered region.

[0160] It should be explained that the "neighborhood noise pixel set" refers to a collection of multiple neighborhood noise pixels. Neighborhood noise pixels refer to all scattered noise pixels within a certain radius (e.g., within a 5×5 pixel window) centered on the central noise pixel. For example, taking the central noise pixel (100, 200) as the center, the 25 pixels within the surrounding 5×5 window are considered as the neighborhood noise pixel set. The "bilateral filter weight set" refers to a collection of multiple bilateral filter weights. Each bilateral filter weight indicates the importance of a particular neighborhood noise pixel in updating the grayscale value of the central noise pixel. This bilateral filter weight comprehensively considers the grayscale difference weight between the neighborhood noise pixel and the central noise pixel, as well as the spatial distance weight. The "weighted filtered grayscale value" refers to the grayscale value obtained after grayscale weighting calculation. The specific method of grayscale weighting calculation is as follows: multiply the grayscale value of each neighborhood noise pixel by its corresponding bilateral filter weight, then sum all the products, and finally divide by the sum of all bilateral filter weights. The result is the weighted filtered grayscale value. The central filtered pixel refers to the central noise pixel after grayscale value update. Grayscale value update means using the weighted filtered grayscale value as the grayscale value of the central noise pixel. The above-mentioned updating the sliding scattering noise region using the central filtered pixel set means using the central filtered pixel set as the pixel value at each position within the sliding scattering noise region.

[0161] In detail, the step of calculating the filtering weights for the central noise pixel based on the neighborhood noise pixel set to obtain a bilateral filtering weight set includes:

[0162] The average gradient of the sliding scattering noise region is calculated based on the set of scattering noise pixels to obtain the average region gradient.

[0163] Obtain the set of scattered gray values ​​corresponding to the set of scattered noise pixels, and use the set of scattered gray values ​​to calculate the gray value variance of the region.

[0164] For each neighboring noisy pixel in the neighborhood noise pixel set, perform the following operation:

[0165] The grayscale difference weight and pixel space weight between the neighboring noise pixels and the center noise pixels are calculated using the average regional gradient and regional grayscale variance.

[0166] The mean of the grayscale difference weight and the pixel space weight is calculated to obtain the bilateral filter weight;

[0167] The bilateral filter weights corresponding to each neighboring noisy pixel are summed to obtain the bilateral filter weight set.

[0168] Optionally, the specific methods for calculating the gray-level difference weights and pixel spatial weights of the neighboring noisy pixels using the average regional gradient and regional gray-level variance include:

[0169] Obtain the neighborhood grayscale value of the neighboring noise pixel and the center grayscale value of the center noise pixel, respectively.

[0170] The grayscale difference weight is calculated based on the neighborhood grayscale value, the center grayscale value, and the average regional gradient.

[0171] Obtain the neighboring pixel coordinates of the neighboring noise pixel and the center pixel coordinates of the center noise pixel, respectively;

[0172] Pixel spatial weights are calculated based on the coordinates of neighboring pixels, the coordinates of the center pixel, and the grayscale variance of the region.

[0173] It should be explained that the average region gradient refers to the average gradient magnitude of all pixels within the sliding scattering noise region, and the region grayscale variance refers to the variance of the grayscale values ​​of all pixels within the sliding scattering noise region. The grayscale difference weight is a value representing the degree of grayscale similarity between neighboring noise pixels and the central noise pixel. The larger the grayscale difference weight, the closer the grayscale values ​​of the neighboring noise pixels and the central noise pixel are, and the higher the importance of the neighboring noise pixel to the central noise pixel. The pixel spatial weight is a value representing the degree of spatial proximity between neighboring noise pixels and the central noise pixel. The larger the pixel spatial weight, the closer the neighboring noise pixels and the central noise pixel are. The neighboring grayscale value refers to the grayscale value of the neighboring pixel. The central grayscale value refers to the grayscale value of the central noise pixel. The above grayscale difference weights are calculated as follows:

[0174] ;

[0175] in, Indicates the weight of grayscale difference. This represents an exponential function with the natural constant as its base. Indicates the center gray value. Represents the grayscale value of the neighborhood, where, The larger the value, the greater the grayscale difference between the neighboring noise pixels and the center noise pixel. The smaller. Indicates the average regional gradient. This represents the grayscale adjustment factor, used to control the adjustment strength of the average region gradient on the grayscale difference weight decay rate. It can be set through cross-validation. The larger the value, the more complex the texture of the sliding scattering noise region. The larger the value, the more significant the difference. The neighboring pixel coordinates refer to the coordinates of the neighboring noise pixels in the scattered noise ray image. The center pixel coordinates refer to the coordinates of the center noise pixel in the scattered noise ray image. The pixel spatial weight is calculated as follows:

[0176] ;

[0177] in, Represents pixel spatial weights, Represents the geometric distance between two coordinates. Indicates the center pixel coordinates. Represents the coordinates of neighboring pixels. This represents the distance adjustment factor, which controls the adjustment strength of the spatial weight decay rate by the regional gray variance. This distance adjustment factor can be set through cross-validation. This represents the variance of gray levels in the region. (The above...) The larger the value, the greater the spatial distance between the neighboring noise pixels and the center noise pixel. The smaller, The larger the value, the higher the noise level in the sliding scattering noise region. The better this is for noise reduction. It should be noted that before calculating the mean of the grayscale difference weights and pixel space weights, both need to be normalized. This is a standard process, and its details will not be elaborated here.

[0178] S5. Identify the current crystal diameter expansion status based on the target ray image to obtain the original crystal diameter expansion data, which includes: diameter expansion timestamp and original crystal radius.

[0179] It is clear that the original crystal diameter expansion data refers to the combination of the diameter expansion timestamp and the original crystal radius.

[0180] In detail, the step of identifying the current crystal diameter expansion state based on the target ray image to obtain the original crystal diameter expansion data includes:

[0181] Edge detection is performed based on the target ray image to obtain the expanded crystal contour, and the coordinates of the contour center in the expanded crystal contour are identified.

[0182] The contour coordinates of the expanded crystal are uniformly extracted to obtain the contour edge coordinate set;

[0183] Based on the coordinates of the contour center, the distance to each contour edge coordinate in the contour edge coordinate set is calculated to obtain the contour edge distance set.

[0184] The original crystal radius is obtained by averaging the distance set of the contour edges;

[0185] Obtain the expansion timestamp corresponding to the original X-ray image, merge the expansion timestamp and the original crystal radius to obtain the original crystal expansion data.

[0186] It should be explained that the expanded diameter crystal contour refers to the boundary line between the silicon carbide crystal and its surrounding environment (such as the inner wall of the graphite crucible or the cavity) in the target ray image. It presents as a closed curve that is approximately circular or elliptical. Edge detection can be performed using the Canny edge detection algorithm. For example, Canny edge detection is performed on the target ray image, where the low threshold of the Canny operator is set to 30 and the high threshold is set to 90, resulting in a binary edge image. The edge pixel coordinates of the crystal region in the binary edge image are (512, 300), (514, 302), ..., (510, 698), totaling 1500 points. These points constitute the expanded diameter crystal contour. The contour center coordinates refer to the coordinates of the geometric center of the expanded diameter crystal contour. For example, by performing least-squares circle fitting on the contour edge coordinate set, the center coordinates are obtained as (511, 499), that is, the contour center coordinates are (511, 499). The contour edge coordinate set refers to a collection of multiple contour edge coordinates. A contour edge coordinate refers to a coordinate point on the expanded crystal contour obtained after uniform extraction of contour coordinates. Uniform extraction of contour coordinates refers to sampling along the expanded crystal contour at equal angular intervals (e.g., every 1 degree) or equal arc length intervals to extract a fixed number of coordinate points, which are the contour edge coordinates. The contour edge distance set refers to a collection of multiple contour edge distances. A contour edge distance refers to the geometric distance between the contour center coordinates and a specific contour edge coordinate. The original crystal radius refers to the average value of the contour edge distance set. The expanded crystal timestamp refers to the timestamp when the original ray image was acquired.

[0187] S6. Summarize the original crystal expansion data corresponding to each original X-ray image to obtain the original crystal expansion dataset. Based on the original crystal expansion dataset, fit the expansion rate to obtain the current expansion rate.

[0188] It is clear that the current diameter expansion rate refers to the growth rate of the radius of the silicon carbide crystal to be monitored at the current growth stage. The specific method for fitting the diameter expansion rate is as follows: using the diameter expansion timestamp as the independent variable and the original crystal radius as the dependent variable, the least squares method is used to perform linear regression on each original crystal diameter expansion data in the original crystal diameter expansion dataset to obtain a regression line. The slope of the regression line is the current diameter expansion rate. If there is a non-linear change in the diameter expansion rate during the growth process, piecewise linear fitting can also be used, and the slope of the most recent fitted segment can be taken as the current diameter expansion rate.

[0189] S7. Adjust the expansion temperature of the crystal expansion equipment according to the current expansion speed and the optimal expansion speed to obtain the adjusted expansion equipment and complete the expansion of the silicon carbide crystal.

[0190] It should be explained that the aforementioned diameter expansion equipment refers to the crystal diameter expansion equipment after diameter expansion temperature adjustment. The specific method for adjusting the diameter expansion temperature of the crystal diameter expansion equipment based on the current diameter expansion speed and the optimal diameter expansion speed is as follows: a proportional-integral-derivative (PID) controller is used, taking the deviation between the current diameter expansion speed and the optimal diameter expansion speed as input, calculating the temperature adjustment amount, and superimposing it on the current growth temperature setpoint, thereby making the diameter expansion speed approach the optimal diameter expansion speed.

[0191] To address the problems described in the background section, this invention first optimizes the diameter expansion rate of the original modification scheme to obtain the optimal expansion rate. This step involves optimizing the diameter expansion rate of the original modification scheme, combining defect detection results with curve fitting and weighting within a discretized speed range to ultimately determine the optimal expansion rate. Compared to the traditional method of setting a fixed expansion rate based on experience, this method dynamically balances crystal quality and growth efficiency, avoiding excessively high speeds that lead to increased edge polycrystalline material or inclusions, and also preventing excessively low speeds that reduce production efficiency. Furthermore, this scheme performs adaptive noise filtering on the original X-ray image to obtain the target X-ray image. This step, through adaptive noise filtering of the original X-ray image, deconstructs the image into two parts: quantum noise and scattering noise, and employs targeted filtering strategies for each, overcoming the limitations of traditional methods. Traditional Gaussian or median filtering suffers from limitations such as inability to distinguish noise types and tendency to blur crystal edges. This invention effectively suppresses high-frequency quantum noise and low-frequency Compton scattering noise while preserving the crystal outline to the maximum extent, providing a high-quality image foundation for subsequent accurate crystal radius measurement. Finally, the crystal expansion equipment's expansion temperature is adjusted based on the current and optimal expansion speeds, resulting in an regulated expansion device. This step uses a PID controller to regulate the crystal expansion equipment temperature based on the deviation between the current and optimal expansion speeds, automatically bringing the expansion speed closer to the optimal value. Compared to the traditional method of manual temperature adjustment, this method achieves closed-loop feedback control, enabling real-time correction of deviations during the growth process and ensuring that the expansion process always operates under optimal conditions, improving the stability and consistency of the expansion process. Therefore, this invention can improve crystal quality and growth efficiency during the expansion process and reduce process fluctuations caused by speed mismatch.

[0192] like Figure 2 The diagram shown is a functional block diagram of a silicon carbide crystal diameter expansion system based on surface modification provided in an embodiment of the present invention.

[0193] The surface-modified silicon carbide crystal diameter expansion system 100 of this invention can be installed in an electronic device. Depending on the functions implemented, the surface-modified silicon carbide crystal diameter expansion system 100 may include a silicon carbide crystal modification module 101, a X-ray image acquisition module 102, a crystal data extraction module 103, and a diameter expansion speed optimization module 104. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and which are stored in the memory of the electronic device.

[0194] The silicon carbide crystal modification module 101 is used to optimize the expansion rate of the pre-acquired original modification scheme to obtain the optimal expansion rate, and to expand the pre-acquired crystal to be expanded according to the original modification scheme and the pre-constructed crystal expansion device to obtain the silicon carbide crystal to be monitored. The crystal expansion device includes an X-ray device.

[0195] The X-ray image acquisition module 102 is used to monitor the image data of the silicon carbide crystal to be monitored using X-ray equipment, obtain the original X-ray image set, extract the original X-ray images sequentially from the original X-ray image set, and perform noise adaptive filtering on the extracted original X-ray images to obtain the target X-ray image.

[0196] The crystal data extraction module 103 is used to identify the current crystal diameter expansion state based on the target ray image and obtain the original crystal diameter expansion data. The original crystal diameter expansion data includes: diameter expansion timestamp and original crystal radius. The original crystal diameter expansion data corresponding to each original ray image is summarized to obtain the original crystal diameter expansion dataset. Based on the original crystal diameter expansion dataset, the diameter expansion speed is fitted to obtain the current diameter expansion speed.

[0197] The expansion speed optimization module 104 is used to adjust the expansion temperature of the crystal expansion equipment according to the current expansion speed and the optimal expansion speed, so as to obtain an adjustable expansion equipment.

[0198] In detail, the modules in the surface-modified silicon carbide crystal diameter expansion system 100 described in this embodiment of the invention employ the same methods as described above during use. Figure 1 The method described herein is the same as the silicon carbide crystal diameter expansion method based on surface modification, and can produce the same technical effect, so it will not be repeated here.

[0199] like Figure 3 The diagram shown is a schematic representation of an electronic device that implements a silicon carbide crystal diameter expansion method based on surface modification, according to an embodiment of the present invention.

[0200] The electronic device 1 may include a processor 10, a memory 11 and a bus 12, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as a method program for expanding the diameter of silicon carbide crystals based on surface modification.

[0201] The memory 11 includes at least one type of readable storage medium, such as flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 1, such as the portable hard drive of the electronic device 1. In other embodiments, the memory 11 can be an external storage device of the electronic device 1, such as a plug-in portable hard drive, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device 1. Furthermore, the memory 11 includes both internal storage units and external storage devices of the electronic device 1. The memory 11 can be used not only to store application software and various types of data installed on the electronic device 1, such as the code of a silicon carbide crystal diameter expansion method program based on surface modification, but also to temporarily store data that has been output or will be output.

[0202] In some embodiments, the processor 10 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 10 is the control unit of the electronic device, connecting various components of the entire electronic device through various interfaces and lines. It executes programs or modules stored in the memory 11 (e.g., a program for a surface-modified silicon carbide crystal diameter expansion method), and calls data stored in the memory 11 to perform various functions of the electronic device 1 and process data.

[0203] The bus 12 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus 12 can be divided into an address bus, a data bus, a control bus, etc. The bus 12 is configured to realize the connection and communication between the memory 11 and at least one processor 10, etc.

[0204] Figure 3 Only electronic devices with components are shown; it will be understood by those skilled in the art that... Figure 3 The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0205] For example, although not shown, the electronic device 1 may also include a power supply (such as a battery) to power the various components. Preferably, the power supply can be logically connected to the at least one processor 10 through a power management device, thereby enabling functions such as charging management, discharging management, and power consumption management. The power supply may also include one or more DC or AC power supplies, recharging devices, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device 1 may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0206] Furthermore, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a Wi-Fi interface, a Bluetooth interface, etc.), which is typically used to establish communication connections between the electronic device 1 and other electronic devices.

[0207] Optionally, the electronic device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), and optionally, a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen, etc. The display may also be appropriately referred to as a screen or display unit, used to display information processed in the electronic device 1 and to display a visual user interface.

[0208] The program for the silicon carbide crystal diameter expansion method based on surface modification, stored in the memory 11 of the electronic device 1, is a combination of multiple instructions. When run in the processor 10, it can achieve the following:

[0209] The diameter expansion rate of the pre-obtained original modification scheme is optimized to obtain the optimal diameter expansion rate;

[0210] The crystal to be expanded is subjected to crystal expansion according to the original modification scheme and the pre-constructed crystal expansion equipment to obtain the silicon carbide crystal to be monitored. The crystal expansion equipment includes an X-ray device.

[0211] X-ray equipment was used to monitor the image data of the silicon carbide crystal under monitoring, and the original X-ray image set was obtained.

[0212] The original ray images are extracted sequentially from the original ray image set, and the extracted original ray images are subjected to noise adaptive filtering to obtain the target ray image;

[0213] The current crystal diameter expansion status is identified based on the target ray image to obtain the original crystal diameter expansion data, which includes: expansion timestamp and original crystal radius;

[0214] The original crystal diameter expansion data corresponding to each original X-ray image are summarized to obtain the original crystal diameter expansion dataset. Based on the original crystal diameter expansion dataset, the diameter expansion rate is fitted to obtain the current diameter expansion rate.

[0215] The crystal expansion equipment is adjusted by regulating the expansion temperature based on the current expansion rate and the optimal expansion rate, thus obtaining an adjustable expansion equipment and completing the silicon carbide crystal expansion.

[0216] Specifically, the processor 10's implementation method for the above instructions can be found in [reference needed]. Figures 1 to 3 The descriptions of the relevant steps in the corresponding embodiments are not repeated here.

[0217] Furthermore, if the modules / units integrated in the electronic device 1 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0218] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor of an electronic device, can perform the following:

[0219] The diameter expansion rate of the pre-obtained original modification scheme is optimized to obtain the optimal diameter expansion rate;

[0220] The crystal to be expanded is subjected to crystal expansion according to the original modification scheme and the pre-constructed crystal expansion equipment to obtain the silicon carbide crystal to be monitored. The crystal expansion equipment includes an X-ray device.

[0221] X-ray equipment was used to monitor the image data of the silicon carbide crystal under monitoring, and the original X-ray image set was obtained.

[0222] The original ray images are extracted sequentially from the original ray image set, and the extracted original ray images are subjected to noise adaptive filtering to obtain the target ray image;

[0223] The current crystal diameter expansion status is identified based on the target ray image to obtain the original crystal diameter expansion data, which includes: expansion timestamp and original crystal radius;

[0224] The original crystal diameter expansion data corresponding to each original X-ray image are summarized to obtain the original crystal diameter expansion dataset. Based on the original crystal diameter expansion dataset, the diameter expansion rate is fitted to obtain the current diameter expansion rate.

[0225] The crystal expansion equipment is adjusted by regulating the expansion temperature based on the current expansion rate and the optimal expansion rate, thus obtaining an adjustable expansion equipment and completing the silicon carbide crystal expansion.

[0226] In the embodiments provided by this invention, it should be understood that the disclosed devices, systems, and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative, and actual implementations may have other classification methods.

[0227] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0228] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0229] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0230] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for increasing the diameter of silicon carbide crystals based on surface modification, characterized in that, The method includes: The diameter expansion rate of the pre-obtained original modification scheme is optimized to obtain the optimal diameter expansion rate; The crystal to be expanded is subjected to crystal expansion according to the original modification scheme and the pre-constructed crystal expansion equipment to obtain the silicon carbide crystal to be monitored. The crystal expansion equipment includes an X-ray device. X-ray equipment was used to monitor the image data of the silicon carbide crystal under monitoring, and the original X-ray image set was obtained. The original ray images are extracted sequentially from the original ray image set, and the extracted original ray images are subjected to noise adaptive filtering to obtain the target ray image; The current crystal diameter expansion status is identified based on the target ray image to obtain the original crystal diameter expansion data, which includes: expansion timestamp and original crystal radius; The original crystal diameter expansion data corresponding to each original X-ray image are summarized to obtain the original crystal diameter expansion dataset. Based on the original crystal diameter expansion dataset, the diameter expansion rate is fitted to obtain the current diameter expansion rate. The crystal expansion equipment is adjusted by regulating the expansion temperature based on the current expansion rate and the optimal expansion rate, thus obtaining an adjustable expansion equipment and completing the silicon carbide crystal expansion.

2. The method for increasing the diameter of silicon carbide crystals based on surface modification as described in claim 1, characterized in that, The step of optimizing the diameter expansion rate of the pre-obtained original modification scheme to obtain the optimal diameter expansion rate includes: The preset expansion speed range is discretized to obtain a discrete expansion speed set; Discrete diameter expansion velocities are extracted sequentially from the discrete diameter expansion velocity set. Based on the original modification scheme and the extracted discrete diameter expansion velocities, the diameter of the pre-obtained experimental silicon carbide crystal is expanded to obtain the experimental diameter-expanded crystal. Defect detection was performed on the experimental expanded-diameter crystal to obtain the experimental defect values; The test defect values ​​corresponding to each discrete diameter expansion rate are summarized to obtain the test defect value set; The optimal expansion speed is determined within the expansion speed range based on the set of experimental defect values.

3. The method for increasing the diameter of silicon carbide crystals based on surface modification as described in claim 2, characterized in that, The step of performing noise adaptive filtering on the extracted original ray image to obtain the target ray image includes: Global gradient data is calculated on the extracted original ray image to obtain the global gradient mean and global noise variance; The original ray image is divided into local windows to obtain multiple windowed ray images; Noise deconstruction is performed using multiple windowed ray images, global gradient mean, and global noise variance to obtain quantum noise ray images and scattered noise ray images; Quantum noise filtering is applied to the quantum noise ray image to obtain a quantum filtered ray image; The scattered noise ray image is filtered to obtain the scattered noise filtered ray image; The target ray image is obtained by synthesizing the quantum-filtered ray image and the scattered ray image.

4. The method for increasing the diameter of silicon carbide crystals based on surface modification as described in claim 3, characterized in that, The method of using multiple windowed ray images, global gradient mean, and global noise variance to perform noise deconstruction to obtain quantum noise ray images and scattered noise ray images includes: Perform the following operation on each of the multiple windowed ray images: Window gradient data calculation is performed on the windowed ray image to obtain the local gradient mean and local noise variance; The noise decision threshold is calculated based on the global gradient mean, global noise variance, local gradient mean, and local noise variance. The windowed ray image is labeled according to the noise decision threshold to obtain the labeled ray image; Adaptive wavelet decomposition is performed on the marked ray image to obtain multiple original wavelet decomposed images; Identify high-frequency and low-frequency wavelet decomposition image groups in multiple original wavelet decomposition images; Noise reconstruction is performed on the high-frequency wavelet decomposition image group to obtain a quantum noise ray image, and noise reconstruction is performed on the low-frequency wavelet decomposition image group to obtain a scattered noise ray image.

5. The method for increasing the diameter of silicon carbide crystals based on surface modification as described in claim 4, characterized in that, The process of performing quantum noise filtering on the quantum noise ray image to obtain a quantum filtered ray image includes: Poisson deviation is calculated on quantum noise ray images to obtain the global Poisson deviation. Multiple sliding quantum noise regions are obtained by sliding the quantum noise ray image through a preset sliding window. Sliding quantum noise regions are extracted sequentially from multiple sliding quantum noise regions, and Poisson deviation is calculated for the extracted sliding quantum noise regions to obtain the local Poisson deviation. Anscombe transform is performed on the extracted sliding quantum noise region based on the global and local Poisson deviations to obtain the sliding Gaussian noise region. By summing up the sliding Gaussian noise regions corresponding to each sliding quantum noise region, multiple sliding Gaussian noise regions are obtained. A global transformed ray image is obtained by merging multiple sliding Gaussian noise regions. Gaussian filtering is applied to the globally transformed ray image to obtain a globally filtered ray image; An inverse Anscombe transform is performed on the global filtered ray image to obtain the quantum filtered ray image.

6. The method for increasing the diameter of silicon carbide crystals based on surface modification as described in claim 5, characterized in that, The process of filtering the scattered noise ray image to obtain a filtered ray image includes: Multiple sliding noise regions are obtained by sliding the scattering noise ray image using a sliding window. For each of the multiple sliding scattering noise regions, the following operation is performed: The set of scattered noise pixels in the sliding scattering noise region is identified, and bilateral filtering is performed using the set of scattered noise pixels to obtain the sliding filtering region. By summarizing the sliding filter regions, multiple sliding filter regions are obtained, and a scattering filter ray image is generated based on these multiple sliding filter regions.

7. The method for increasing the diameter of silicon carbide crystals based on surface modification as described in claim 6, characterized in that, The method of using the set of scattered noise pixels to perform bilateral filtering to obtain a sliding filtering region includes: Scattered noise pixels are extracted sequentially from the set of scattered noise pixels, and the extracted scattered noise pixels are recorded as the center noise pixel. Based on the central noise pixel, identify the neighborhood noise pixel set within the scattered noise pixel set; The filtering weights of the central noise pixel are calculated based on the neighborhood noise pixel set to obtain the bilateral filtering weight set; The weighted gray value is obtained by using the bilateral filter weight set and the neighborhood noise pixel set to perform gray-level weighted calculation. The grayscale value of the center noise pixel is updated using the weighted filtered grayscale value to obtain the center filtered pixel; The central filtered pixels corresponding to each central noise pixel are summed to obtain the central filtered pixel set; The sliding scattering noise region is updated using the central filtered pixel set to obtain the sliding filtered region.

8. The method for increasing the diameter of silicon carbide crystals based on surface modification as described in claim 7, characterized in that, The step of calculating the filtering weights for the center noise pixel based on the neighborhood noise pixel set to obtain a bilateral filtering weight set includes: The average gradient of the sliding scattering noise region is calculated based on the set of scattering noise pixels to obtain the average region gradient. Obtain the set of scattered gray values ​​corresponding to the set of scattered noise pixels, and use the set of scattered gray values ​​to calculate the gray value variance of the region. For each neighboring noisy pixel in the neighborhood noise pixel set, perform the following operation: The grayscale difference weight and pixel space weight between the neighboring noise pixels and the center noise pixels are calculated using the average regional gradient and regional grayscale variance. The mean of the grayscale difference weight and the pixel space weight is calculated to obtain the bilateral filter weight; The bilateral filter weights corresponding to each neighboring noisy pixel are summed to obtain the bilateral filter weight set.

9. The method for increasing the diameter of silicon carbide crystals based on surface modification as described in claim 8, characterized in that, The step of identifying the current crystal diameter expansion state based on the target ray image to obtain the original crystal diameter expansion data includes: Edge detection is performed based on the target ray image to obtain the expanded crystal contour, and the coordinates of the contour center in the expanded crystal contour are identified. The contour coordinates of the expanded crystal are uniformly extracted to obtain the contour edge coordinate set; Based on the coordinates of the contour center, the distance to each contour edge coordinate in the contour edge coordinate set is calculated to obtain the contour edge distance set. The original crystal radius is obtained by averaging the distance set of the contour edges; Obtain the expansion timestamp corresponding to the original X-ray image, merge the expansion timestamp and the original crystal radius to obtain the original crystal expansion data.

10. A silicon carbide crystal diameter expansion system based on surface modification, characterized in that, The system includes: The silicon carbide crystal modification module is used to optimize the expansion rate of the pre-acquired original modification scheme to obtain the optimal expansion rate. Based on the original modification scheme and the pre-constructed crystal expansion equipment, the module expands the diameter of the pre-acquired crystal to be expanded to obtain the silicon carbide crystal to be monitored. The crystal expansion equipment includes an X-ray device. The X-ray image acquisition module is used to monitor the image data of the silicon carbide crystal under monitoring using X-ray equipment, obtain the original X-ray image set, extract the original X-ray images sequentially from the original X-ray image set, and perform noise adaptive filtering on the extracted original X-ray images to obtain the target X-ray image. The crystal data extraction module is used to identify the current crystal diameter expansion state based on the target ray image and obtain the original crystal diameter expansion data. The original crystal diameter expansion data includes: diameter expansion timestamp and original crystal radius. The original crystal diameter expansion data corresponding to each original ray image is summarized to obtain the original crystal diameter expansion dataset. Based on the original crystal diameter expansion dataset, the diameter expansion rate is fitted to obtain the current diameter expansion rate. The diameter expansion speed optimization module is used to adjust the diameter expansion temperature of the crystal diameter expansion equipment according to the current diameter expansion speed and the optimal diameter expansion speed, so as to obtain the adjustable diameter expansion equipment.

Images