Method, equipment and medium for improving real-time online measurement accuracy of crystal growth thickness

Abstract

The invention relates to a method and equipment for improving real-time online measurement accuracy of crystal growth thickness and a medium. The method comprises the following steps: acquiring a first image and a second image shot in each growth period in a crystal growth process; determining a first online measurement thickness according to the first image, calibrating a second calibration frame and a second crystal center position in a first second image of each growth cycle, and adjusting image visual attributes of all the second images in the corresponding growth cycle according to the second calibration frame and the second crystal center position to obtain a second target image; determining a longitudinal pixel size according to a preset table surface diameter, a central pixel of the second target image and an included angle between the second view angle direction and the first view angle direction; and determining a target crystal image covered by the second calibration frame in the second target image in the second visual angle direction, and determining a second online measurement thickness of each crystal in the second visual angle direction according to the pixel value, the included angle and the longitudinal pixel size of the pixel in the target crystal image.

Description

Technical Field

[0001] The present disclosure relates to the technical field of crystal growth, and particularly to a method, device and medium for improving the accuracy of real-time online measurement of crystal growth thickness. Background Art

[0002] Crystal growth is a key process in the manufacturing of semiconductors, optics and laser devices, and its thickness control directly affects the performance of the final product. However, the existing measurement of crystal growth thickness cannot be carried out in real-time online, and mostly adopts the off-line measurement method, which requires interrupting the growth process for sampling and detection, resulting in low process efficiency and inability to adjust the growth parameters in real-time. This reduces the efficiency of crystal production. Moreover, when performing real-time online measurement, it focuses on the measurement of the crystal growth thickness in a single direction. During the production process, useless crystals generated at the edge of the molybdenum stage will affect the crystal thickness that really needs to be measured, resulting in a decrease in the accuracy of crystal growth thickness measurement. Summary of the Invention

[0003] The purpose of the present invention is to provide a method, device and medium for improving the accuracy of real-time online measurement of crystal growth thickness, aiming to solve the technical problems existing in the prior art that the crystal growth thickness cannot be measured in real-time online, resulting in low crystal production efficiency and useless crystals generated at the edge of the molybdenum stage and the like will affect the crystal thickness that really needs to be measured, resulting in a decrease in the accuracy of crystal growth thickness measurement.

[0004] To achieve the above purpose, in the first aspect of the embodiments of the present disclosure, a method for improving the accuracy of real-time online measurement of crystal growth thickness is provided, and the method includes:

[0005] Obtain a first image captured by a first camera and a second image captured by a second camera during each growth cycle in the crystal growth process. The first camera is correspondingly arranged with a first observation window in the first viewing direction of the vacuum reaction chamber, and the second camera is correspondingly arranged with a second observation window in the second viewing direction of the vacuum reaction chamber. There is an included angle between the first viewing direction and the second viewing direction;

[0006] Determine a first online measurement thickness according to the first image, and in the first second image of each growth cycle, calibrate a second calibration frame and a second crystal center position, and adjust the image visual attributes of all the second images in the corresponding growth cycle according to the second calibration frame and the second crystal center position to obtain a second target image;

[0007] Determine a longitudinal pixel size according to the preset table diameter of the molybdenum stage, the central pixel of the second target image, and the included angle of the second viewing direction relative to the first viewing direction. The longitudinal pixel size is used to represent the size represented by 1 pixel in the longitudinal direction in the first viewing direction;

[0008] Determine the target crystal image covered by the second calibration box in the second target image in the second viewing direction, where the target crystal image is composed of pixels in the second target image whose pixel values are within the primary color value range;

[0009] Determine the second on-line measured thickness of each crystal in the second viewing direction according to the pixel values of the pixels in the target crystal image, the included angle, and the longitudinal pixel size.

[0010] In a second aspect of the embodiments of the present disclosure, a crystal growth apparatus is provided. The crystal growth apparatus includes: a control terminal, a vacuum reaction chamber, a microwave energy providing unit, a process gas providing unit, an electric proportional valve, a vacuum pump, a first camera, and a second camera;

[0011] The vacuum reaction chamber is used to provide a crystal growth environment. A molybdenum stage is arranged in the vacuum reaction chamber, and the molybdenum stage is used to carry the grown crystal. A first observation window and a second observation window are arranged in different viewing directions of the vacuum reaction chamber. The first camera is arranged at the first observation window, and the second camera is arranged at the second observation window;

[0012] The first camera takes a first image of the inside of the vacuum reaction chamber from a first viewing direction through the first observation window, and the second camera takes a second image of the inside of the vacuum reaction chamber from a second viewing direction through the second observation window;

[0013] The microwave energy providing unit and the process gas providing unit are respectively connected to the vacuum reaction chamber through pipelines, and the vacuum pump is connected to the vacuum reaction chamber through a pipeline. The electric proportional valve is arranged on the pipeline connecting the vacuum pump and the vacuum reaction chamber;

[0014] The control terminal is respectively communicatively connected to the vacuum reaction chamber, the microwave energy providing unit, the process gas providing unit, the electric proportional valve, the vacuum pump, the first camera, and the second camera, and is used to control the growth of the crystal, obtain the first image and the second image, and execute the method according to any one of the first aspect.

[0015] In a third aspect of the embodiments of the present disclosure, an electronic device is provided, including:

[0016] A memory on which a computer program is stored;

[0017] A processor for executing the computer program in the memory to implement the steps of the method according to any one of the first aspect.

[0018] In a fourth aspect of the embodiments of the present disclosure, there is provided a computer-readable storage medium having stored thereon a computer program, which when executed by a processor, implements the steps of the method according to any one of the first aspect.

[0019] The present invention provides a method, an apparatus, and a medium for improving the accuracy of real-time online measurement of crystal growth thickness. Compared with the prior art, the following beneficial effects are achieved:

[0020] By arranging a first camera and a second camera in different viewing angle directions of a vacuum reaction chamber, image information from multiple angles during the crystal growth process can be obtained simultaneously. This arrangement helps to more comprehensively and accurately understand the growth state of the crystal. Determining the first online measurement thickness using the first image provides a basis for the preliminary thickness evaluation of the crystal. Then, by calibrating a second calibration frame and the position of the second crystal center in the second image and adjusting the image visual attributes to obtain a second target image, the accuracy and efficiency of image processing are improved. Combining the preset table diameter of the molybdenum table, the central pixel of the second target image, and the viewing angle, the longitudinal pixel size can be accurately calculated. In the second viewing angle direction, by screening pixels with pixel values within the primary color value range, the target crystal image can be accurately identified, avoiding interference from the background or other impurities. According to the pixel values of the target crystal image, the viewing angle, and the longitudinal pixel size, the second online measurement thickness of each crystal in the second viewing angle direction can be accurately calculated. This online measurement method not only improves the real-time performance of the measurement but also ensures the accuracy of the measurement.

[0021] In this way, by real-time monitoring and accurately measuring the thickness of the crystal, abnormal situations during the crystal growth process, such as abnormal growth rate, crystal defects, etc., can be detected in a timely manner, so as to adjust the growth parameters in a timely manner, optimize the crystal growth process, and improve the quality and yield of the crystal. Through means such as multi-dimensional monitoring, accurate measurement and calibration, efficient target crystal image recognition, and accurate online thickness measurement, real-time monitoring and precise control of the crystal growth process are achieved. Without interrupting the crystal growth process, the growth parameters can be dynamically adjusted, improving the process efficiency. Thus, the crystal growth process can be monitored in real time, the growth parameters can be dynamically adjusted, and the process efficiency is significantly improved.

[0022] Other features and advantages of the present disclosure will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The drawings are used to provide a further understanding of the present disclosure and constitute a part of the specification. They are used together with the following specific implementation to explain the present disclosure, but do not constitute a limitation to the present disclosure. In the drawings:

[0024] Figure 1 A flowchart of a method for improving the accuracy of real-time online measurement of crystal growth thickness according to an embodiment of the specification.

[0025] Figure 2 It is a schematic diagram of a crystal growth device shown according to an embodiment of the specification.

[0026] Figure 3 It is a schematic diagram of a first calibration frame and the position of the center of the first crystal shown according to an embodiment of the specification.

[0027] Figure 4 It is a schematic diagram of a preset fixed search area and a crystal in a molybdenum stage shown according to an embodiment of the specification.

[0028] Figure 5 It is a schematic diagram of a second calibration frame and the position of the center of the second crystal shown according to an embodiment of the specification.

[0029] Figure 6 It is a block diagram of a control terminal shown according to an embodiment of the specification. Detailed implementation manners

[0030] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0031] The following will describe in detail the specific implementation manners of the present disclosure in conjunction with the accompanying drawings. It should be understood that the specific implementation manners described herein are only used to illustrate and explain the present disclosure, and are not used to limit the present disclosure.

[0032] The present disclosure provides a method for improving the accuracy of real-time online measurement of crystal growth thickness. Figure 1 It is a flowchart of a method for improving the accuracy of real-time online measurement of crystal growth thickness shown according to an embodiment. The method for improving the accuracy of real-time online measurement of crystal growth thickness specifically includes the following steps:

[0033] In step S11, a first image captured by a first camera and a second image captured by a second camera in each growth cycle during the crystal growth process are obtained. The first camera is correspondingly arranged with a first observation window in the first viewing direction of the vacuum reaction chamber, the second camera is correspondingly arranged with a second observation window in the second viewing direction of the vacuum reaction chamber, and there is an included angle between the first viewing direction and the second viewing direction.

[0034] In the embodiments of the present disclosure, in the crystal growth experiment, by precisely controlling the reaction conditions (such as temperature, pressure, gas composition, etc.) in the vacuum reaction chamber, the raw materials are gradually crystallized in the vacuum reaction chamber. The first camera monitors this process continuously through the first observation window, and the second camera monitors this process continuously through the second observation window, and takes images at the end of each growth cycle.

[0035] Among them, the distance between the camera and the observation window can be determined according to the camera focal length, the size of the observation window, and the required image resolution, usually ensuring that the camera can clearly capture the entire crystal area. The distance between the observation window and the crystal: It is determined by the design of the vacuum reaction chamber to ensure that the camera's field of view can cover the entire crystal growth area.

[0036] In one embodiment, as shown in Figure 2 shown, the first observation window can be set horizontally, and the second observation window can be set at an angle to the horizontal direction. Therefore, the first viewing direction can be the horizontal direction, and thus the first camera can take the first image in each growth cycle during the crystal growth process from the horizontal direction. The second viewing direction has an angle with the horizontal direction and can take the second image in each growth cycle during the crystal growth process from an oblique direction.

[0037] In step S12, determine the first online measurement thickness according to the first image, and in the first second image of each growth cycle, calibrate the second calibration frame and the second crystal center position, and adjust the image visual attributes of all the second images in the corresponding growth cycle according to the second calibration frame and the second crystal center position to obtain the second target image;

[0038] In the embodiments of the present disclosure, determining the first online measurement thickness according to the first image may include:

[0039] In the first first image of each growth cycle, with the goal of covering at least all the crystals in the first viewing direction, construct a first calibration frame, and calibrate the center position of any crystal within the first calibration frame in the first viewing direction as the first crystal center position;

[0040] Among them, the calibration frame is a rectangular or circular area set artificially in the image for positioning and analyzing the target object. The first crystal center position: It is the geometric center of any crystal determined by the image processing algorithm within the calibration frame.

[0041] In the embodiments of the present disclosure, in the first image of each growth cycle, use image processing software or algorithms to automatically identify and construct a calibration frame covering all the crystals. Then, through image processing techniques such as edge detection and centroid calculation, determine the center position of any crystal within the calibration frame.

[0042] Among them, the calibration box size is dynamically adjusted according to the distribution and size of the crystals in the image to ensure that all crystals are covered. The distance between the center position of the crystal and the edge of the calibration box is automatically calculated by the image processing algorithm and is used for subsequent analysis and monitoring.

[0043] See Figure 3 As shown, all the grown crystals are framed by the calibration box of the gray rectangle in the first long first image, and the center position of the crystal on the left side within the first calibration box is calibrated as the first crystal center position ( Figure 3 indicated by the black "cross" in). Among them, the crystal area to be measured needs to be manually located, the first calibration box is calibrated, and a crystal center position is calibrated so that the crystal is more prominent (manual operation is only required in the first picture during one growth cycle, and subsequent calibration is not required).

[0044] Obtain the three primary color information of the pixels at the first crystal center position corresponding to each growth cycle during the crystal growth process, and determine the primary color value range of each primary color according to the three primary color information of the pixels at the first crystal center position;

[0045] Among them, the three primary color information refers to the intensity values of the three basic colors, red (R), green (G), and blue (B), of the pixels at the first crystal center position in the first image. The primary color value range is the value range of the intensity of each primary color.

[0046] In the embodiments of the present disclosure, for the pixels at the first crystal center position in each growth cycle, their three primary color information is obtained. Then, according to this information, the value range of each primary color is determined. At the determined crystal center position, the RGB values of the pixel are read, and the RGB values of all the center position pixels during the entire growth cycle are statistically analyzed, so as to determine the value range of each primary color.

[0047] According to the first calibration box and the primary color value ranges of each primary color, determine the parameter values of the image visual attributes corresponding to the first image, obtain the target parameter values of the image visual attributes, and adjust the image visual attributes of all the first images corresponding to the growth cycle according to the target parameter values to obtain the first target image;

[0048] Among them, the image visual attributes refer to the attributes of the image such as brightness, contrast, and saturation, which affect the visual effect of the image. The target parameter values are the ideal values of the image visual attributes determined according to specific requirements or standards.

[0049] In the embodiments of the present disclosure, according to the calibration frame and the original color value range, the parameter values of the image visual attributes (such as brightness, contrast, etc.) are determined, and then these attributes of all images within the corresponding growth period are adjusted to obtain the target image. According to the determined original color value range, the attributes such as brightness and contrast of the image are adjusted to make the crystal clearer and more prominent in the image.

[0050] According to the preset thickness of the molybdenum stage, the target region image in the preset fixed search region in the first target image, and the number of crystal pixels between the upper edge and the lower edge of each column of the crystal within the first calibration frame in the first target image, the first online measurement thickness of each crystal measured online in the first viewing direction is obtained.

[0051] Among them, molybdenum stage: a platform used to support the crystal during crystal growth, usually made of high-temperature resistant materials such as molybdenum. Target region image: a specific region specified in the image for analysis or measurement. Number of crystal pixels: the number of pixels representing the height of the crystal in the image. First online measurement thickness: the thickness of the crystal measured in real time by image processing technology during crystal growth.

[0052] In the embodiments of the present disclosure, using the known thickness of the molybdenum stage, the target region image, and the number of pixels between the upper edge and the lower edge of the crystal, the online measurement thickness of the crystal is calculated through a proportional relationship. For example, before the start of crystal growth, a measuring tool (caliper) is used to measure the thickness S t (unit: mm). For example, assume the thickness of the molybdenum stage is 1 mm. In the target region image, it is measured that there are 500 pixels between the upper edge and the lower edge of a certain crystal. At the same time, the actual size represented by each pixel in the image is known. Through proportional calculation, the online measurement thickness of this crystal can be obtained.

[0053] The above technical solution ensures the accuracy and comprehensiveness of the measurement area by constructing the first calibration frame and calibrating the position of the center of the first crystal. At the same time, by analyzing the three primary color information of the pixels at the center position of the first crystal, the influence of image noise and light change is eliminated, and the image quality is improved. According to the value of the target parameter, the visual attributes of the image are adjusted to further optimize the image clarity and contrast, ensuring the accuracy of crystal edge detection and adapting to the complex lighting conditions and crystal surface reflection characteristics in the vacuum reaction chamber by analyzing the three primary color information and adjusting the visual attributes of the image. By adjusting the visual attributes of the image, the clarity, contrast, and color balance of the image are optimized, ensuring the clear visibility of the crystal edge and details, and effectively overcoming the problem of image quality degradation in high-temperature and high-pressure environments. Not only is the measurement accuracy of the crystal growth thickness significantly improved, with the error controlled at the sub-micron level, meeting the requirements of high-precision crystal growth, but also the first image is obtained in real time during each growth cycle, and through automated image processing and data calculation, the real-time online measurement of the crystal thickness is realized. There is no need to interrupt the crystal growth process, and the growth parameters can be dynamically adjusted to improve the process efficiency. Thus, the crystal growth process can be monitored in real time, and the growth parameters can be dynamically adjusted, significantly improving the process efficiency.

[0054] In a possible implementation manner, obtaining the first online measurement thickness of each crystal in the first viewing direction according to the preset thickness of the molybdenum stage, the target region image in the preset fixed search region in the first target image, and the number of crystal pixels between the upper edge and the lower edge of each column of the crystal in the crystal within the first calibration frame in the first target image includes:

[0055] Determine the vertical pixel size according to the preset thickness of the molybdenum stage and the target region image in the preset fixed search region in the first target image, where the vertical pixel size is used to represent the size represented by 1 pixel in the vertical direction;

[0056] Among them, the vertical pixel size: In the image, the actual physical size represented by each pixel in the vertical direction. This size usually needs to be determined through calibration or known physical sizes.

[0057] In the embodiments of the present disclosure, using the known thickness of the molybdenum stage and its pixel representation in the first target image, the actual size represented by each pixel in the vertical direction can be calculated. This calculation process usually involves a proportional relationship, that is, the ratio of the actual thickness of the molybdenum stage to the number of pixels in the image.

[0058] For example: Assume that the actual thickness of the molybdenum stage is 1 mm, and in the first target image, the height of the molybdenum stage occupies 100 pixels. Then, the vertical pixel size can be calculated as 1 mm / 100 pixels = 0.01 mm / pixel. This means that in the first target image, each pixel in the vertical direction represents an actual size of 0.01 mm.

[0059] Calculate the number of crystal pixels between the upper edge and the lower edge of each column of the crystal constructed by the continuous pixels within the first calibration box in the first target image;

[0060] Among them, the number of crystal pixels: In the image, it represents the number of pixels of the crystal height. This quantity can be measured by image processing techniques, usually involving steps such as edge detection and contour extraction.

[0061] In the embodiments of the present disclosure, in the first target image, for each column of crystals within the first calibration box, the upper edge and the lower edge of the crystal are determined by image processing techniques (such as edge detection algorithms). Then, calculate the number of pixels between these two edges, that is, the number of crystal pixels.

[0062] For example, in the first target image, select a column of clear crystals for analysis. Use the edge detection algorithm to identify the upper edge and the lower edge of the crystal, and calculate the number of pixels between these two edges. Suppose the calculation result is 200 pixels, then the number of crystal pixels of this column of crystals is 200.

[0063] According to the number of crystal pixels and the vertical pixel size, obtain the first on-line measurement thickness of each crystal in the first viewing direction.

[0064] Among them, the first on-line measurement thickness: During the crystal growth process, it is the thickness of the crystal measured in real time by image processing techniques. This thickness is calculated based on the number of pixels in the image and the vertical pixel size.

[0065] In the embodiments of the present disclosure, using the determined vertical pixel size and the calculated number of crystal pixels, the first on-line measurement thickness of the crystal can be calculated by a simple multiplication operation. That is, multiply the number of crystal pixels by the vertical pixel size to obtain the actual thickness of the crystal.

[0066] For example, assume that the determined vertical pixel size is 0.01 mm / pixel and the calculated number of crystal pixels is 200. Then, the first on-line measurement thickness of the crystal can be calculated as 200 pixels * 0.01 mm / pixel = 2 mm. This means that in the first viewing direction, the thickness of this column of crystals is approximately 2 mm.

[0067] The above technical solution can realize the on-line measurement of the crystal thickness during the crystal growth process. This method has the advantages of non-contact, real-time, accuracy, etc.

[0068] In a possible implementation manner, in step S151, determining the vertical pixel size according to the preset thickness of the molybdenum stage and the target region image in the preset fixed search region in the first target image includes:

[0069] Traverse and calculate the pixel value difference between adjacent pixels in the vertical direction in the target region image from top to bottom to obtain a first pixel value difference, and determine the first pixel whose first pixel value difference exceeds the preset difference range for the first time as the upper edge pixel belonging to the upper edge of the molybdenum stage;

[0070] Among them, the pixel value difference is the pixel value difference between adjacent pixels in the image. This difference can be used to detect edges or mutation points in the image. The preset difference range is a threshold range set during the pixel value difference calculation. When the difference exceeds this range, it is considered that there is a significant edge or mutation between adjacent pixels. The upper edge pixel is the pixel in the image that represents the upper edge of the object. This pixel is usually located in the transition region between the object and the background, and its pixel value difference will exceed the preset difference range.

[0071] In the embodiments of the present disclosure, traverse the pixels in the target region image from top to bottom, and calculate the pixel value difference between each pixel and its adjacent pixel below. When this difference exceeds the preset difference range for the first time, it is considered that this pixel belongs to the upper edge of the molybdenum stage, that is, the upper edge pixel.

[0072] For example, assume that there is an obvious brightness difference between the upper edge of the molybdenum stage and the background in the target region image. Traverse the image from top to bottom. When the brightness difference between a certain pixel and its adjacent pixel below is calculated to exceed the preset brightness difference range for the first time, it can be considered that this pixel is the upper edge pixel of the molybdenum stage.

[0073] Traverse and calculate the pixel value difference between adjacent pixels in the vertical direction in the target region image from bottom to top to obtain a second pixel value difference, and determine the second pixel whose second pixel value difference exceeds the preset difference range for the first time as the lower edge pixel belonging to the lower edge of the molybdenum stage;

[0074] Among them, the lower edge pixel is the pixel in the image that represents the lower edge of the object. Similar to the upper edge pixel, this pixel is also located in the transition region between the object and the background, but its position is below the object.

[0075] In the embodiments of the present disclosure, pixels in the target region image are traversed from bottom to top, and the pixel value difference between each pixel and its adjacent pixel above is calculated. When this difference first exceeds the preset difference range, it is considered that this pixel belongs to the lower edge of the molybdenum stage, that is, the lower edge pixel.

[0076] For example, when traversing the image from bottom to top, when the luminance difference between a certain pixel and its adjacent pixel above first exceeds the preset luminance difference range, it can be considered that this pixel is the lower edge pixel of the molybdenum stage.

[0077] Determine the number of vertical pixels according to the number of pixels between the upper edge pixel and the lower edge pixel;

[0078] In the embodiments of the present disclosure, the number of vertical pixels is the number of pixels representing the height of an object in the image. This number can be obtained by calculating the number of pixels between the upper edge pixel and the lower edge pixel. After determining the upper edge pixel and the lower edge pixel, by calculating the number of pixels between these two pixels, the number of vertical pixels of the molybdenum stage in the image can be obtained. For example, assume that the index of the upper edge pixel is i and the index of the lower edge pixel is j (and i < j). Then, the number of vertical pixels of the molybdenum stage is j - i.

[0079] Determine the vertical pixel size according to the ratio between the preset thickness of the molybdenum stage and the vertical pixel size.

[0080] In the embodiments of the present disclosure, using the known thickness (preset thickness) of the molybdenum stage and its number of vertical pixels in the image, the vertical pixel size can be calculated through a proportional relationship. That is, divide the thickness of the molybdenum stage by the number of vertical pixels to obtain the actual physical size represented by each pixel.

[0081] See Figure 4 As shown, in Figure 4 the A frame area is a fixed search area. In this area, the RGB values of all pixels are compared; compare two by two from top to bottom, and a group of pixels with a difference greater than the set value (range 50 - 200) and in the same vertical position are determined as the upper edge of the molybdenum stage; (such as Figure 4 the upper edge of the B frame area in Figure 4 ); compare two by two from bottom to top, and a group of pixels with a difference greater than the set value (range 50 - 200) and in the same vertical position are determined as the lower edge of the molybdenum stage; (such as the lower edge of the B frame area in

[0082] Calculation of the number of pixels in the vertical direction of the molybdenum stage: Accumulate the number of pixels between the upper edge and the lower edge of a single column on the molybdenum stage, which is the number of pixels Sm in the vertical direction of the molybdenum stage; Input the true thickness dimension of the molybdenum stage. Before the start of crystal growth, use a measuring tool (caliper) to measure the thickness dimension St (unit: mm) of the molybdenum stage; Calculation of the vertical pixel size: The vertical pixel size (unit: mm) Sz = St / Sm.

[0083] Within the calibration area ( Figure 3 the medium gray box), determine the pixels that meet the RGB upper and lower limit ranges and are connected as crystals (such as Figure 4 the C and D box areas in are crystals): Calculation of the number of vertical pixels of the crystal. For the same crystal, calculate the number of pixels between the upper edge and the lower edge of each column of the crystal separately, and the maximum number of pixels is the number of pixels Sc in the vertical direction of this crystal. Calculation of the crystal thickness: The crystal thickness in the horizontal direction (unit: mm) Sh = Sc x Sz.

[0084] The above technical solution can accurately determine the vertical pixel size. This size is the basis for subsequent calculation of parameters such as crystal thickness, improving the accuracy of measurement.

[0085] In a possible implementation manner, in step S14, the determining the parameter value of the image visual attribute corresponding to the first image according to the first calibration box and the primary color value range of each primary color, and obtaining the target parameter value of the image visual attribute includes:

[0086] Starting from the preset initial value of the image visual attribute, adjust the parameter value of the image visual attribute step by step according to the preset step size;

[0087] Among them, the preset initial value is the initial setting value of the image visual attribute parameter before the start of adjustment. The preset step size is the change amount of the parameter value each time the image visual attribute parameter is adjusted.

[0088] In the embodiments of the present disclosure, starting from the preset initial value of the image visual attribute, increase or decrease the parameter value of this attribute step by step according to the preset step size. This process is usually implemented through image processing algorithms or software and can be automated.

[0089] For example, assume that the brightness attribute of the image needs to be adjusted, the preset initial value is 100 (representing a certain intermediate value of the brightness level), and the preset step size is 5. Then, starting from 100, increase or decrease by 5 each time, and try different brightness values in turn.

[0090] After each adjustment of the parameter value, count the number of pixels of the three primary colors within the range covered by the first calibration box that are in the primary color value range of each primary color;

[0091] In the embodiments of the present disclosure, after each adjustment of the value of the image visual attribute parameter, the tristimulus values of each pixel within the coverage of the first calibration box are recalculated, and the number of pixels whose values fall within their respective primary color value ranges is counted. This process generally involves steps such as image segmentation, color space conversion, and pixel counting.

[0092] For example, assume that the primary color value range of red is set as [100, 150], that of green is [50, 100], and that of blue is [75, 125]. After each adjustment of the brightness, the RGB values of each pixel within the first calibration box are calculated, and the number of pixels that satisfy the above-mentioned red, green, and blue value ranges is counted.

[0093] Determine the value of the parameter corresponding to the largest number as the target parameter value of the image visual attribute.

[0094] Among them, the target parameter value is the parameter value that, in a series of parameter value attempts, makes the number of pixels that satisfy a specific condition (such as the number of pixels whose tristimulus values are within the primary color value ranges) the largest.

[0095] In the embodiments of the present disclosure, after counting the number of pixels whose tristimulus values are within their respective primary color value ranges under all attempted parameter values, select the parameter value with the largest number as the target parameter value of the image visual attribute. This process generally involves comparison and selection operations.

[0096] For example, the brightness is cyclically adjusted from 0% to 100% (increasing by 1% each time). Each time the brightness is adjusted, the number of pixels within the calibration area that meet the RGB upper and lower limits is calculated. The brightness value with the largest number is the optimal brightness. After the calculation is completed, the brightness is adjusted to the optimal brightness. According to the method of brightness adjustment, the contrast, saturation, clarity, sharpness, and color temperature are adjusted to the optimal values to obtain the first target image.

[0097] It is possible to determine the value of the image visual attribute parameter that makes the number of pixels whose tristimulus values within the coverage of the first calibration box are within their respective primary color value ranges the largest, that is, the target parameter value of the image visual attribute, so as to facilitate the identification of the grown crystal from the image, thereby improving the accuracy of online thickness measurement.

[0098] In a possible implementation manner, the image visual attribute includes at least one of the following: brightness, contrast, saturation, clarity, sharpness, color temperature.

[0099] In a possible implementation manner, in step S13, the determining the primary color value range of each primary color according to the tristimulus information of the pixels at the center position of the first crystal includes:

[0100] Determine the primary color limit values of each primary color according to the values of each primary color represented in the three - primary - color information of the pixel at the first crystal center position and a preset expansion value. Specifically, add the preset expansion value to the value of each primary color to obtain the upper limit value of the primary color limit value, and subtract the preset expansion value from the value of each primary color to obtain the lower limit value of the primary color limit value.

[0101] Among them, in an image, red, green, and blue (RGB) are the three basic colors that make up the image color, and each color has a value range (usually 0 - 255). The preset expansion value is to ensure the accuracy and robustness of color recognition. When calculating the primary color value range, usually according to the color characteristics of the target object, an extended range is set for its three - primary - color values, and the size of this range is the preset expansion value. The primary color limit value is the color value obtained by adding or subtracting the preset expansion value based on the three - primary - color information, which are respectively called the upper limit value and the lower limit value.

[0102] In the embodiments of the present disclosure, for each pixel at the first crystal center position, its RGB value is extracted. Then, based on these RGB values, the preset expansion value is added and subtracted respectively to obtain the upper limit value and the lower limit value of each primary color. This process helps to consider the error range of color recognition and improve the recognition accuracy.

[0103] For example, assume that the RGB value of the pixel at the first crystal center position is (100, 150, 200), and the preset expansion value is 20. Then, the upper limit value of red is 120 (100 + 20), and the lower limit value is 80 (100 - 20); the upper limit value of green is 170 (150 + 20), and the lower limit value is 130 (150 - 20); the upper limit value of blue is 220 (200 + 20), and the lower limit value is 180 (200 - 20).

[0104] Determine the primary color value range of each primary color according to the upper limit value and the lower limit value of each primary color.

[0105] In the embodiments of the present disclosure, according to the calculated upper limit value and lower limit value of each primary color, the value range of each primary color can be determined. This range will be used for subsequent color matching to identify pixels or regions similar to the color of the first crystal. For example, continuing the above example, the value range of red is [80, 120], the value range of green is [130, 170], and the value range of blue is [180, 220]. These ranges will be used to find regions similar to the color of the first crystal in the image.

[0106] In summary, according to the three primary color information of the pixels at the center position of the first crystal and in combination with a preset expansion value, the value range of each primary color can be determined. This process has wide application value in the fields of image processing, machine vision, color recognition, etc., especially when precise identification of specific color objects is required.

[0107] For example, in the first first image, taking the values of each primary color of the three primary color information represented by the image at the calibrated center position of the first crystal, that is, the RGB values of the pixels as a reference, calculate the upper and lower limits of RGB:

[0108] R+ = Ra + A; R- = Ra - A;

[0109] G+ = Ga + A; G- = Ga - A;

[0110] B+ = Ba + A; B- = Ba - A;

[0111] A is the preset expansion value. R+: the upper limit value of the red primary color, R-: the lower limit value of the red primary color, and the primary color value range of the red primary color is: [Ra - A, Ra + A]. Similarly, G+: the upper limit value of the green primary color, G-: the lower limit value of the green primary color, and the primary color value range of the green primary color is: [Ga - A, Ga + A]; B+: the upper limit value of the blue primary color, B-: the lower limit value of the blue primary color, and the primary color value range of the blue - red primary color is: [Ba - A, Ba + A].

[0112] Similarly, in the first second image of each growth cycle, manually or automatically calibrate the second calibration frame and the center position of the second crystal. Then, according to this calibration information, adjust the image visual attributes of all second images within the corresponding growth cycle to improve the image quality and the accuracy of subsequent processing.

[0113] In step S13, according to the preset table diameter of the molybdenum table, the central pixel of the second target image, and the angle between the second viewing direction and the first viewing direction, determine the longitudinal pixel size, where the longitudinal pixel size is used to represent the size represented by 1 pixel in the longitudinal direction under the first viewing direction;

[0114] Among them, the preset table diameter of the molybdenum table: the known diameter of the molybdenum table (the platform for carrying the crystal). The longitudinal pixel size: the actual size represented by 1 pixel in the longitudinal direction in the image under the first viewing direction.

[0115] In the embodiments of the present disclosure, the longitudinal pixel size is determined through geometric calculation by using the preset tabletop diameter of the molybdenum table, the central pixel of the second target image, and the included angle between the second viewing direction and the first viewing direction. This size is a key parameter for subsequent calculation of the second online measurement thickness. Assume that the preset tabletop diameter of the molybdenum table is 100 mm, the central pixel coordinates of the second target image are (x, y), and the included angle between the second viewing direction and the first viewing direction is 45 degrees. Through geometric calculation, the longitudinal pixel size can be obtained as 0.1 mm / pixel (assumed value).

[0116] In step S14, under the second viewing direction, determine the target crystal image covered by the second calibration frame in the second target image, where the target crystal image is composed of pixels in the second target image whose pixel values are within the primary color value range;

[0117] In the embodiments of the present disclosure, under the second viewing direction, pixels belonging to the crystal are screened out from the second target image according to the primary color value range to form a target crystal image. This image is used for subsequent thickness measurement. Assume that the primary color value range is [150, 255] (assumed value for grayscale images), then the pixels in the second target image with pixel values within this range are considered to be part of the crystal and form the target crystal image.

[0118] In step S15, determine the second online measurement thickness of each crystal under the second viewing direction according to the pixel values of the pixels in the target crystal image, the included angle, and the longitudinal pixel size.

[0119] In the embodiments of the present disclosure, according to the pixel values of the pixels in the target crystal image, the included angle, and the longitudinal pixel size, the second online measurement thickness of each crystal under the second viewing direction is determined through geometric calculation. This thickness value reflects the actual thickness of the crystal in the current growth cycle. Assume that the pixel value of a certain pixel in the target crystal image is 200 (gray value), the projection length of this pixel in the second viewing direction is L (obtained through geometric calculation), and the longitudinal pixel size is 0.1 mm / pixel. Then the second online measurement thickness can be obtained by multiplying L by the reciprocal of the longitudinal pixel size (i.e., L / 0.1 mm). In this way, the first online measurement thickness and the second online measurement thickness can be combined as the online measurement thickness of the final crystal.

[0120] The above technical solution can simultaneously obtain multi-angle image information during crystal growth by arranging a first camera and a second camera in different viewing angle directions of the vacuum reaction chamber. This arrangement helps to more comprehensively and accurately understand the growth state of the crystal. Using the first image to determine the first online measured thickness provides a basis for the preliminary thickness evaluation of the crystal. Then, by calibrating the second calibration frame and the second crystal center position in the second image and adjusting the image visual attributes to obtain the second target image, the accuracy and efficiency of image processing are improved. Combining the preset table diameter of the molybdenum table, the central pixel of the second target image, and the viewing angle, the longitudinal pixel size can be accurately calculated. In the second viewing angle direction, by screening the pixels with pixel values within the primary color value range, the target crystal image can be accurately identified, avoiding interference from the background or other impurities. According to the pixel value of the target crystal image, the viewing angle, and the longitudinal pixel size, the second online measured thickness of each crystal in the second viewing angle direction can be accurately calculated. This online measurement method not only improves the real-time performance of the measurement but also ensures the accuracy of the measurement.

[0121] In this way, by real-time monitoring and accurately measuring the thickness of the crystal, abnormal situations during crystal growth, such as abnormal growth rate, crystal defects, etc., can be detected in a timely manner, so as to adjust the growth parameters in a timely manner, optimize the crystal growth process, and improve the quality and yield of the crystal. Through means such as multi-dimensional monitoring, accurate measurement and calibration, efficient target crystal image recognition, and accurate online thickness measurement, real-time monitoring and accurate control of the crystal growth process are achieved. Without interrupting the crystal growth process, the growth parameters can be dynamically adjusted to improve the process efficiency. Thus, the crystal growth process can be monitored in real time, and the growth parameters can be dynamically adjusted, significantly improving the process efficiency.

[0122] In a possible implementation manner, in step S13, the determining the longitudinal pixel size according to the preset table diameter of the molybdenum table, the central pixel of the second target image, and the angle between the second viewing angle direction and the first viewing angle direction includes:

[0123] In step S131, taking the central pixel of the second target image as the reference pixel, determine the long diameter and short diameter of the molybdenum table under the viewing angle of the second viewing angle direction;

[0124] Among them, the reference pixel is a reference point for determining the size and position of an object in the image, which is the central pixel of the second target image here. The long diameter is the projection length of the molybdenum table in the longest direction under the second viewing angle direction. The short diameter is the projection length of the molybdenum table in the shortest direction under the second viewing angle direction.

[0125] In the embodiments of the present disclosure, in the second target image, with the central pixel as the reference, the contour of the molybdenum stage is determined through image processing techniques (such as edge detection, contour extraction, etc.), and its major diameter and minor diameter in the second viewing direction are measured.

[0126] In step S132, according to the ratio between the preset stage diameter of the molybdenum stage and the major diameter, the horizontal pixel size in the horizontal direction in the second viewing direction is determined, and the horizontal pixel size is used to represent the size represented by 1 pixel in the horizontal direction in the second viewing direction;

[0127] Among them, the horizontal pixel size is the actual size represented by 1 pixel in the horizontal direction in the image in the second viewing direction.

[0128] In the embodiments of the present disclosure, given the preset stage diameter (true diameter) of the molybdenum stage and its major diameter in the second target image, the horizontal pixel size is calculated through ratio calculation. The formula is: horizontal pixel size = preset stage diameter / major diameter (in pixels).

[0129] In step S133, based on trigonometric functions, according to the horizontal pixel size and the angle between the second viewing direction and the first viewing direction, the longitudinal deformation diameter of the molybdenum stage is determined;

[0130] Among them, the longitudinal deformation diameter is the diameter after longitudinal deformation of the molybdenum stage relative to its true size in the second viewing direction due to the viewing angle change. Based on trigonometric functions (such as sine, cosine, etc.), using the known horizontal pixel size and viewing angle, the longitudinal deformation diameter of the molybdenum stage in the second viewing direction is calculated.

[0131] Among them, before the crystal growth starts, using a measuring tool (caliper), the diameter size B of the molybdenum stage surface is measured d (unit: mm), the horizontal pixel size (unit: mm) B m = B d / B; for the calculation of the vertical pixel size, before the crystal growth starts, query the cavity processing drawing to confirm that the installation angle of the second camera (angle camera) relative to the bottom horizontal plane of the cavity is ∠C. Due to the angle shooting, the longitudinal diameter of the molybdenum stage surface will be deformed to (unit: mm) A d = Sin∠C x B d .

[0132] In step S134, according to the ratio between the longitudinal deformation diameter and the minor diameter, the vertical pixel size is determined.

[0133] In the embodiments of the present disclosure, the preset tabletop diameter (true diameter) of the molybdenum table and its short diameter in the second target image (the longitudinal projection length after perspective distortion) are known, and the longitudinal pixel size is obtained through ratio calculation. However, it should be noted here that since the short diameter is measured in the second viewing direction and the longitudinal deformed diameter has been calculated through trigonometric functions, this information is used to inversely calculate the longitudinal pixel size in the first viewing direction (or relative to a specific direction).

[0134] In a possible implementation manner, in step S131, determining the long diameter and short diameter of the molybdenum table in the viewing angle of the second viewing direction with the central pixel of the second target image as the reference pixel includes:

[0135] In step S1311, with the central pixel of the second image as the first reference pixel, calculate the pixel value differences between adjacent pixels in each row in turn along any one side in the first viewing direction to obtain the third pixel value difference until the third pixel value difference exceeds the preset difference range for the first time in each row, and stop the calculation of that row;

[0136] Among them, the third pixel value difference is the pixel value difference between adjacent pixels in the row-by-row calculation. Preset difference range: a set threshold used to determine whether there is a significant pixel value change between adjacent pixels, usually determined based on the edge characteristics of the object in the image. In the embodiments of the present disclosure, starting from the central pixel of the second target image, scan row by row along the first viewing direction (or its perpendicular direction, specifically depending on how "any one side" is defined), and calculate the pixel value differences between adjacent pixels in each row. This difference reflects the sharpness of the object edge in the image. When the difference exceeds the preset difference range, it is considered that the edge of the object has been encountered.

[0137] In step S1312, determine the row-edge pixel belonging to the lateral edge of the molybdenum table among the adjacent pixels where the third pixel value difference exceeds the preset difference range for the first time in each row, and is close to the first reference pixel;

[0138] Among them, the row-edge pixel is the pixel among the adjacent pixels where the pixel value difference exceeds the preset difference range for the first time in any row and is close to the first reference pixel (i.e., the central pixel), which marks the lateral edge position of the object in that row.

[0139] In the embodiments of the present disclosure, for each row, find the adjacent pixel pair where the pixel value difference exceeds the preset difference range for the first time, and select the pixel close to the central pixel as the row-edge pixel. These pixels together constitute the lateral edge contour of the molybdenum table in the second viewing direction.

[0140] In step S1313, in the direction perpendicular to the first viewing direction, the two row-wise edge pixels with the largest distance are used as the second reference pixels, and the distance between the second reference pixels is determined as the major diameter of the molybdenum stage under the viewing angle of the second viewing direction;

[0141] Among them, the second reference pixels are the two row-wise edge pixels with the largest distance in the direction perpendicular to the first viewing direction, and they are respectively located at both ends of the lateral edge of the molybdenum stage. After determining all the row-wise edge pixels, find the two row-wise edge pixels with the largest distance in the direction perpendicular to the first viewing direction (i.e., both ends of the lateral edge of the molybdenum stage). The distance between these two pixels is the major diameter of the molybdenum stage under the second viewing direction.

[0142] For example, taking the pixel at the center of the picture as the reference, ignoring the yellow frame area and the upper 50% area of the picture, compare pixel by pixel from top to bottom, left and right. The pixel closer to the center with a difference value greater than the set value (range 50 - 200) is determined as the lateral edge B of the molybdenum stage 左(X,Y) 、B 右(X,Y) , and the maximum difference between the two pixels on the left and right is determined as the major diameter B of the molybdenum stage (unit: pixel):

[0143] B = B 右MAX(X) -B 左MAX(X)

[0144] In step S1314, respectively starting from the second reference pixels, construct reference lines parallel to the first viewing direction, and calculate the pixel value differences between adjacent pixels in each column in turn along the direction perpendicular to the second viewing direction until the fourth pixel value difference in each column first exceeds the preset difference range, and then stop the calculation of this column;

[0145] Among them, the fourth pixel value difference: the pixel value difference between adjacent pixels in the column-by-column calculation, which is used to detect the edge change in the column direction. Reference line: a straight line constructed starting from the second reference pixel and parallel to the first viewing direction, which is used to guide the detection of the column-wise edge.

[0146] Starting from the two second reference pixels, construct two reference lines parallel to the first viewing direction respectively. Then, scan column by column along the direction perpendicular to the second viewing direction (i.e., perpendicular to the reference line direction), and calculate the pixel value differences between adjacent pixels in each column. This difference reflects the sharpness of the edge of the object in the column direction in the image.

[0147] In step S1315, the pixel closer to the reference line among the adjacent pixels where the fourth pixel value difference first exceeds the preset difference range in each column is determined as the column-wise edge pixel belonging to the column-wise edge of the molybdenum stage;

[0148] Among them, the columnar edge pixel is the pixel closer to the reference line among adjacent pixels where the pixel value difference first exceeds a preset difference range in each column, and it marks the edge position of the object in that column.

[0149] For each column, find adjacent pixel pairs where the pixel value difference first exceeds the preset difference range, and select the pixel closer to the reference line as the columnar edge pixel. These pixels together constitute the columnar edge contour of the molybdenum stage in the second viewing direction.

[0150] In step S1316, the pixel with the maximum distance from the reference line among the column edge pixels is used as the short diameter pixel, and based on the distance between the short diameter pixel and the second reference pixel in the first viewing direction, the short diameter of the molybdenum stage in the second viewing direction is determined.

[0151] Among them, the short diameter pixel is the pixel with the maximum distance from the reference line among the columnar edge pixels, and it marks the endpoint of the short diameter of the molybdenum stage in the second viewing direction.

[0152] After determining all the columnar edge pixels, find the columnar edge pixel with the maximum distance from the reference line. This pixel is one endpoint of the short diameter. Since the position of the reference line (i.e., the position of the second reference pixel) is known, the short diameter can be determined by calculating the distance between this columnar edge pixel and the corresponding reference line in the first viewing direction.

[0153] For example, taking B 右MAX(Y) and B 左MAX(Y) as the reference line, ignoring the yellow frame area, starting from B 左MAX(X,Y) to B 右MAX(X,Y) compare column by column downward in pairs. The pixel closer to the reference line with a difference greater than the set value (range 50 - 200) is determined as the lower edge A of the molybdenum stage 下(X,Y) , and the edge with the maximum distance from the reference line is determined as the short diameter A of the molybdenum stage (unit: pixel):

[0154] A = (B 左MAX(Y) - A 下MAX(Y) ) x 2

[0155] In a possible implementation manner, in step S15, the determining the second online measurement thickness of each crystal in the second viewing direction according to the pixel values of the pixels in the target crystal image, the included angle, and the longitudinal pixel size includes:

[0156] In step S151, determine the target crystal image covered by the second calibration frame in the second target image in the second viewing direction, where the target crystal image is composed of pixels with pixel values within the primary color value range in the second target image;

[0157] In the embodiments of the present disclosure, in the second target image, a second calibration box is used to determine a region that contains the image of the target crystal. Then, all pixels within this region are traversed, and pixels that make up the target crystal image are filtered out based on whether their pixel values fall within the primary color value range.

[0158] In step S152, based on the pixel values of the pixels in the target crystal image, the upper edge line and the lower edge line of the crystal are determined;

[0159] Among them, the upper edge line of the crystal is the set of pixels at the top edge of the crystal in the target crystal image, which form a line marking the position of the upper surface of the crystal. The lower edge line of the crystal is the set of pixels at the bottom edge of the crystal in the target crystal image, which also form a line marking the position of the lower surface of the crystal.

[0160] In the embodiments of the present disclosure, after the target crystal image is determined, the upper edge line and the lower edge line of the crystal are determined by analyzing the change in pixel values. This usually involves edge detection algorithms such as Canny edge detection and Sobel operator to identify positions where pixel values change drastically in the image, thereby determining the edges.

[0161] In step S153, based on the difference in the number of pixels between the upper edge line of the crystal and the lower edge line of the crystal, the number of side pixels of the crystal is determined;

[0162] Among them, the number of side pixels of the crystal is the number of pixels along the side direction of the crystal between the upper edge line of the crystal and the lower edge line of the crystal. This number reflects the actual size of the crystal in the second viewing direction (representation in the image).

[0163] After determining the upper edge line and the lower edge line of the crystal, the difference in the number of pixels between them can be calculated (or more accurately, the number of pixels on the shortest distance along the side direction of the crystal). This number is the number of side pixels of the crystal.

[0164] In step S154, based on the number of side pixels of the crystal and the longitudinal pixel size, the thickness of each crystal during deformation in the second viewing direction is determined, and based on the thickness during deformation and the included angle, the second online measurement thickness of each crystal in the second viewing direction is determined.

[0165] Among them, the longitudinal pixel size (unit: mm) A m = A d / A. Crystal thickness calculation (result during deformation): Thickness of the crystal during deformation (unit: mm) C m = C x A m ; True thickness of the crystal (unit: mm) Dm = C m / Cos∠C。

[0166] The S obtained through the above calculations h and D m are the real-time values of the crystal growth thickness (unit: mm), which are output to the MPCVD control system, and the control system records and displays the data output.

[0167] In a possible implementation manner, in step S152, determining the upper edge line and the lower edge line of the crystal according to the pixel values of the pixels in the target crystal image includes:

[0168] In step S1521, the straight line where the lowest edge covered by the second calibration frame in the second target image is located is determined as the lower edge line of the crystal, and the lowest edge of the target crystal image coincides with the lowest edge of the second calibration frame;

[0169] Among them, since the lowest edge of the target crystal image coincides with the lowest edge of the second calibration frame, the straight line where the lowest edge covered by the second calibration frame in the second target image can be directly determined as the lower edge line of the crystal. This assumption is based on the prior knowledge that the crystal is vertically placed in the image and the bottom is aligned with the bottom of the calibration frame.

[0170] In step S1522, for each column of pixels in the target crystal image, the pixel value difference between two adjacent pixels above and below is calculated in sequence to obtain the fifth pixel value difference, and the calculation of this column is stopped until the fifth pixel value difference exceeds the preset difference range for the first time in each column;

[0171] Among them, the fifth pixel value difference: the pixel value difference between two adjacent pixels above and below in each column of the target crystal image, which is used to detect the sharpness of the upper edge of the crystal. For each column of pixels in the target crystal image, the pixel value difference between two adjacent pixels above and below is calculated in sequence until the situation where the pixel value difference exceeds the preset difference range appears for the first time in each column. This preset difference range is used to distinguish the pixel value change between the crystal edge and the background or noise.

[0172] In step S1523, the pixel corresponding to the pixel among the two adjacent pixels above and below where the fifth pixel value difference exceeds the preset difference range for the first time in each column and is close to the pixel located at the center in the target crystal image is determined as the upper edge pixel belonging to the upper edge of the molybdenum stage;

[0173] Among them, the upper edge pixel is the pixel closer to the center of the image among the two adjacent pixels above and below where the pixel value difference exceeds the preset difference range for the first time in each column of the target crystal image, and it is considered to be a part of the upper edge of the crystal.

[0174] For each column, find the pair of adjacent pixels above and below where the difference in pixel values first exceeds the preset difference range, and select the pixel closer to the center of the target crystal image as the upper edge pixel. This selection is based on the prior knowledge that the crystal is centered in the image.

[0175] In step S1524, determine the line where the upper edge pixels with the most identical row coordinates are located as the upper edge line of the crystal.

[0176] Among them, the upper edge line of the crystal is a straight line obtained by fitting all the upper edge pixels, which is used to represent the upper boundary of the crystal in the image. After determining all the upper edge pixels, use a linear fitting algorithm (such as the least squares method) to fit the row coordinates of these pixels, so as to obtain the equation of the upper edge line of the crystal. This fitting process assumes that the upper edge of the crystal is a straight line, and the row coordinates of the upper edge pixels are approximately on this straight line.

[0177] In the embodiment of the present disclosure, within the calibration area, the pixel combinations that meet the RGB upper and lower limit ranges and are connected are determined as crystals. Within the identified crystal area, taking the central pixel of the crystal as a reference, starting from the left edge to the right edge of the crystal, compare two by two column by column downward. The pixel closer to the reference with a difference greater than the set value (range 50 - 200) is determined as the upper edge of the crystal side. The horizontal line with the largest number of upper edges determined at the same X position is the upper edge line C on(Y) of the crystal. The lower edge of the identified crystal area is directly used as the lower edge line C down(Y) of the crystal side. Furthermore, calculate the difference between C on(Y) and C down(Y), which is the number of pixels on the crystal side C (unit: pixel).

[0178] In a possible implementation manner, in step S12, in the first second image of each growth cycle, calibrate the second calibration frame and the second crystal center position, and adjust the image visual attributes of all the second images in the corresponding growth cycle according to the second calibration frame and the second crystal center position to obtain a second target image, including:

[0179] In step S121, in the first second image of each growth cycle, with the goal of covering at least all the crystals in the second viewing direction, construct a second calibration frame, and calibrate the center position of any crystal within the second calibration frame in the second viewing direction as the second crystal center position;

[0180] Among them, in the first and second images of each growth cycle, with the goal of covering at least all the crystals in the second viewing direction, a second calibration box is manually or automatically constructed. Then, within the second calibration box, the center position of any crystal is determined through image processing algorithms (such as the centroid method, ellipse fitting method, etc.) and calibrated as the second crystal center position. This center position can be used for subsequent image processing and visual attribute adjustment.

[0181] See Figure 5 As shown, a rectangular box is marked as the second calibration box, and the center position of the right crystal is used as the second crystal center position.

[0182] In step S122, obtain the three primary color information of the pixels at the corresponding second crystal center position during each growth cycle in the crystal growth process;

[0183] Among them, at the second crystal center position, obtain the three primary color information of the corresponding pixels. This usually involves reading the RGB values of the pixels at this position in the image.

[0184] In step S123, according to the values of each primary color represented in the three primary color information of the pixels at the second crystal center position and a preset expansion value, determine the upper and lower limit values of each primary color. Among them, add the preset expansion value to the value of each primary color to obtain the upper limit value of the primary color limit value, and subtract the preset expansion value from the value of each primary color to obtain the lower limit value of the primary color limit value;

[0185] Among them, the primary color limit value: the limit value obtained by adding or subtracting a preset expansion value based on the primary color value, used to define the value range of the primary color.

[0186] According to the three primary color information of the pixels at the second crystal center position and a preset expansion value (usually determined based on experience or experiments), calculate the upper and lower limit values of each primary color. These limit values will be used to determine the value interval of the primary color. The calculation method is the same as that for the first image, so no exemplary description is given.

[0187] In step S124, according to the upper limit value and the lower limit value of each primary color, determine the primary color value interval of each primary color;

[0188] Similar to the processing of the first image, define the value interval of each primary color according to the determined upper limit value and lower limit value. These intervals will be used to adjust the visual attributes of the image subsequently.

[0189] In step S125, according to the second calibration box and the value ranges of the respective primary colors, determine the parameter values of the image visual attributes corresponding to the second image, obtain the target parameter values of the image visual attributes, and adjust the image visual attributes of all the second images in the corresponding growth period according to the target parameter values to obtain the second target image.

[0190] Among them, image visual attributes: attributes such as the brightness, contrast, and saturation of an image, which determine the visual effect of the image. Target parameter values: The parameter values of the image visual attributes determined according to the second calibration box and the value ranges of the respective primary colors, used to adjust the image to obtain the desired visual effect.

[0191] According to the second calibration box and the value ranges of the respective primary colors, determine the target parameter values of the image visual attributes. Then, use these parameter values to adjust the visual attributes of all the second images in the corresponding growth period to obtain the second target image. This usually involves adjusting attributes such as the brightness, contrast, and saturation of the image.

[0192] In a possible implementation manner, the determining the parameter values of the image visual attributes corresponding to the second image according to the second calibration box and the value ranges of the respective primary colors includes:

[0193] Starting from the preset initial value of the image visual attribute, adjust the parameter values of the image visual attribute step by step according to the preset step size;

[0194] Among them, the preset initial value is the initial setting value of the image visual attribute parameter before the start of adjustment. The preset step size is the change amount of the parameter value each time the image visual attribute parameter is adjusted.

[0195] In the embodiments of the present disclosure, starting from the preset initial value of the image visual attribute, increase or decrease the parameter values of this attribute step by step according to the preset step size. This process is usually implemented through image processing algorithms or software and can be automated.

[0196] For example, assume that it is necessary to adjust the brightness attribute of an image, the preset initial value is 100 (representing a certain intermediate value of the brightness level), and the preset step size is 5. Then, starting from 100, increase or decrease by 5 each time, and try different brightness values in turn.

[0197] After each adjustment of the parameter values, count the number of pixels of the three primary colors within the range covered by the second calibration box that are within the value ranges of the respective primary colors;

[0198] In the embodiments of the present disclosure, after each adjustment of the value of the image visual attribute parameter, the tristimulus values of each pixel within the coverage of the first calibration box are recalculated, and the number of pixels whose values are within their respective primary color value ranges is counted. This process generally involves steps such as image segmentation, color space conversion, and pixel counting.

[0199] For example, assume that the primary color value range of red is set as [100, 150], that of green is [50, 100], and that of blue is [75, 125]. After each adjustment of the brightness, the RGB values of each pixel within the first calibration box are calculated, and the number of pixels that satisfy the above-mentioned red, green, and blue value ranges is counted.

[0200] Determine the value of the parameter corresponding to the largest number as the target parameter value of the image visual attribute.

[0201] Among them, the target parameter value is the parameter value that, in a series of parameter value attempts, makes the number of pixels that satisfy specific conditions (such as the tristimulus values being within the primary color value ranges) the largest.

[0202] In the embodiments of the present disclosure, after counting the number of pixels whose tristimulus values are within their respective primary color value ranges under all attempted parameter values, select the parameter value with the largest number as the target parameter value of the image visual attribute. This process generally involves comparison and selection operations.

[0203] For example, the brightness is cyclically adjusted from 0% to 100% (incrementing by 1% each time). Each time it is adjusted, the number of pixels within the calibration area that meet the RGB upper and lower limit ranges is calculated. The brightness value with the largest number is the optimal brightness. After the calculation is completed, the brightness is adjusted to the optimal brightness. According to the method of brightness adjustment, the contrast, saturation, sharpness, sharpness, and color temperature are adjusted to the optimal values to obtain the first target image.

[0204] It is possible to determine the value of the image visual attribute parameter that makes the number of pixels whose tristimulus values within the coverage of the first calibration box are within their respective primary color value ranges the largest, that is, the target parameter value of the image visual attribute, so as to facilitate the identification of the grown crystal from the image, thereby improving the accuracy of online thickness measurement.

[0205] The embodiments of the present disclosure further provide a crystal growth device. Refer to Figure 2 As shown, the crystal growth device includes: a control terminal, a vacuum reaction chamber, a microwave energy providing unit, a process gas providing unit, an electric proportional valve, a vacuum pump, a first camera, and a second camera;

[0206] The vacuum reaction chamber is used to provide a crystal growth environment. A molybdenum stage is arranged in the vacuum reaction chamber, and the molybdenum stage is used to carry the grown crystal. A first observation window and a second observation window are arranged in different viewing directions of the vacuum reaction chamber. The first camera is arranged at the first observation window, and the second camera is arranged at the second observation window;

[0207] The first camera takes a first image of the inside of the vacuum reaction chamber from a first viewing direction through the first observation window, and the second camera takes a second image of the inside of the vacuum reaction chamber from a second viewing direction through the second observation window;

[0208] The microwave energy supply unit and the process gas supply unit are respectively connected to the vacuum reaction chamber through pipelines. The vacuum pump is connected to the vacuum reaction chamber through a pipeline, and an electric proportional valve is arranged on the pipeline connecting the vacuum pump and the vacuum reaction chamber;

[0209] The control terminal is respectively communicatively connected to the vacuum reaction chamber, the microwave energy supply unit, the process gas supply unit, the electric proportional valve, the vacuum pump, the first camera and the second camera, and is used to control the growth of the crystal, obtain the first image and the second image, and execute the method described in any one of the foregoing embodiments.

[0210] In the embodiments of the present disclosure, the vacuum reaction chamber provides a specific environment for crystal growth. By controlling the air pressure, temperature and atmosphere inside it, the crystal growth conditions can be precisely adjusted. This usually involves using a vacuum pump to reduce the air pressure inside the chamber, and precisely controlling the introduction of process gas through an electric proportional valve, so as to maintain the required growth environment.

[0211] Microwave energy heating: The microwave energy supply unit heats the substances in the vacuum reaction chamber through microwave radiation. Microwave heating has the characteristics of high efficiency and uniformity, and can quickly and precisely control the temperature in the reaction chamber, which is crucial for crystal growth.

[0212] Growth process monitoring: The first camera takes images of the inside of the vacuum reaction chamber from a specific viewing angle through the first observation window, so that the control terminal can monitor the crystal growth state in real time. This monitoring is crucial for timely adjusting the growth parameters, ensuring crystal quality and growth efficiency.

[0213] Automation control: The control terminal, acting as the "brain" of the entire system, is communicatively connected to various components (such as a vacuum reaction chamber, a microwave energy supply unit, a process gas supply unit, an electric proportional valve, a vacuum pump, and a first camera). It is responsible for receiving image data from the camera, analyzing the growth state of the crystal, and adjusting the working parameters of each component according to a preset algorithm or user instructions to achieve automated control of crystal growth.

[0214] In summary, the crystal growth device ensures high-quality crystal growth and improves process efficiency by precisely controlling the environmental parameters (such as air pressure, temperature, and atmosphere) in the vacuum reaction chamber, using microwave energy for efficient heating, real-time online monitoring of the crystal growth state, and achieving automated control of the system.

[0215] The embodiments of the present disclosure also provide a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the steps of the method described in any one of the foregoing embodiments are implemented.

[0216] The embodiments of the present disclosure also provide an electronic device, including:

[0217] A memory, on which a computer program is stored;

[0218] A processor, configured to execute the computer program in the memory to implement the steps of the method described in any one of the foregoing embodiments.

[0219] Figure 6 The shown real-time online measuring device 100 for crystal growth thickness includes: a processor 1001 and a memory 1003. Among them, the processor 1001 and the memory 1003 are connected, such as through a bus 1002. Optionally, the real-time online measuring device 100 for crystal growth thickness may further include a communication component 1004, and the communication component 1004 can be used for data interaction between the device 100 and other devices, such as data sending and / or data receiving, etc. It should be noted that in actual scheduling, the communication component 1004 is not limited to one, and the structure of the real-time online measuring device 100 for crystal growth thickness does not constitute a limitation to the embodiments of the present application.

[0220] The processor 1001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute various exemplary logical blocks, modules, and circuits described in connection with the disclosure of this application. The processor 1001 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0221] The bus 1002 may include a path for transmitting information between the above components. The bus 1002 may be a PCI (Peripheral Component Interconnect) bus, an EISA (Extended Industry Standard Architecture) bus, or the like. The bus 1002 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6 only a thick line is used to represent it herein, but it does not mean that there is only one bus or one type of bus.

[0222] The memory 1003 may be a ROM (Read Only Memory) or other type of static storage device that can store static information and instructions, a RAM (Random Access Memory) or other type of dynamic storage device that can store information and instructions, or it may also be an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory), or other optical disc storage, optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media, other magnetic storage devices, or any other medium that can be used to carry or store program code and can be read by a computer, which is not limited herein.

[0223] The memory 1003 is used to store the program code for implementing the embodiments of the present disclosure, and is controlled by the processor 1001 for execution. The processor 1001 is configured to execute the program code stored in the memory 1003 to implement the steps shown in the foregoing method embodiments for improving the accuracy of real-time on-line measurement of crystal growth thickness.

[0224] Embodiments of the present disclosure further provide a computer-readable storage medium, on which program code is stored. When the program code is executed by a processor, the steps and corresponding contents of the foregoing method embodiments for improving the accuracy of real-time on-line measurement of crystal growth thickness can be implemented.

[0225] The preferred embodiments of the present disclosure have been described in detail above in conjunction with the accompanying drawings. However, the present disclosure is not limited to the specific details in the above embodiments. Within the scope of the technical concept of the present disclosure, these embodiments can be subject to various changes, modifications, substitutions, and variations, and these changes, modifications, substitutions, and variations all fall within the protection scope of the present disclosure.

[0226] In addition, it should be noted that, in the above specific embodiments, the various specific technical features described can be combined in any appropriate manner without conflict, and the same should be regarded as the content disclosed by the present disclosure. To avoid unnecessary repetition, the present disclosure does not separately describe various possible combination manners. The technical scope of this application is not limited to the content in the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A method for improving the accuracy of real-time on-line measurement of crystal growth thickness, characterized in that, The method includes: Obtaining a first image captured by a first camera and a second image captured by a second camera in each growth cycle during crystal growth. The first camera is correspondingly arranged with a first observation window in a first viewing direction of a vacuum reaction chamber, and the second camera is correspondingly arranged with a second observation window in a second viewing direction of the vacuum reaction chamber. There is an included angle between the first viewing direction and the second viewing direction; Determining a first online measurement thickness according to the first image, and calibrating a second calibration frame and a second crystal center position in the first second image of each growth cycle. Then, adjusting the image visual attributes of all the second images in the corresponding growth cycle according to the second calibration frame and the second crystal center position to obtain a second target image; Determining a longitudinal pixel size according to a preset table diameter of a molybdenum stage, the central pixel of the second target image, and the included angle of the second viewing direction relative to the first viewing direction. The longitudinal pixel size is used to represent the size represented by 1 pixel in the longitudinal direction in the first viewing direction; Determining a target crystal image covered by the second calibration frame in the second target image in the second viewing direction, where the target crystal image is composed of pixels with pixel values within the primary color value range in the second target image; Determining a second online measurement thickness of each crystal in the second viewing direction according to the pixel values of the pixels in the target crystal image, the included angle, and the longitudinal pixel size.

2. The method according to claim 1, characterized in that The determining the longitudinal pixel size according to the preset table diameter of the molybdenum stage, the central pixel of the second target image, and the included angle of the second viewing direction relative to the first viewing direction includes: Taking the central pixel of the second target image as a reference pixel to determine the long diameter and short diameter of the molybdenum stage in the viewing angle of the second viewing direction; Determining a horizontal pixel size in the horizontal direction in the second viewing direction according to the ratio between the preset table diameter of the molybdenum stage and the long diameter. The horizontal pixel size is used to represent the size represented by 1 pixel in the horizontal direction in the second viewing direction; Based on trigonometric functions, determining the longitudinal deformation diameter of the molybdenum stage according to the horizontal pixel size and the included angle of the second viewing direction relative to the first viewing direction; Determining the longitudinal pixel size according to the ratio between the longitudinal deformation diameter and the short diameter.

3. The method according to claim 2, characterized in that, The taking the central pixel of the second target image as a reference pixel to determine the long diameter and short diameter of the molybdenum stage in the viewing angle of the second viewing direction includes: Taking the central pixel of the second image as a first reference pixel, calculating the pixel value difference between adjacent pixels in each row in any one side along the first viewing direction to obtain a third pixel value difference until the third pixel value difference exceeds a preset difference range for the first time in each row, and then stopping the calculation of that row; Determining the row-edge pixel belonging to the horizontal edge of the molybdenum stage as the pixel closer to the first reference pixel among the adjacent pixels where the third pixel value difference exceeds the preset difference range for the first time in each row; In the direction perpendicular to the first viewing direction, take the two row-edge pixels with the largest distance as the second reference pixels, and determine the distance between the second reference pixels as the major diameter of the molybdenum stage under the viewing angle in the second viewing direction; Taking the second reference pixels as starting points respectively, construct reference lines parallel to the first viewing direction, and calculate the pixel value differences between adjacent pixels in each column successively along the direction perpendicular to the second viewing direction to obtain the fourth pixel value difference until the fourth pixel value difference in each column first exceeds the preset difference range, and then stop the calculation of this column; Determine the column-edge pixels belonging to the column edge of the molybdenum stage as the adjacent pixels closer to the reference line among the adjacent pixels where the fourth pixel value difference first exceeds the preset difference range in each column; Take the pixel with the largest distance from the reference line among the column-edge pixels as the minor diameter pixel, and determine the minor diameter of the molybdenum stage under the viewing angle in the second viewing direction according to the distance between the minor diameter pixel and the second reference pixel in the direction along the first viewing direction.

4. The method according to claim 1, wherein The determining the second online measurement thickness of each crystal in the second viewing direction according to the pixel values of the pixels in the target crystal image, the included angle, and the longitudinal pixel size includes: Determine the target crystal image covered by the second calibration frame in the second target image in the second viewing direction, where the target crystal image is composed of the pixels with pixel values within the primary color value range in the second target image; Determine the upper crystal edge line and the lower crystal edge line according to the pixel values of the pixels in the target crystal image; Determine the number of crystal side pixels according to the difference in the number of pixels between the upper crystal edge line and the lower crystal edge line; Determine the thickness at the time of deformation of each crystal in the second viewing direction according to the number of crystal side pixels and the longitudinal pixel size, and determine the second online measurement thickness of each crystal in the second viewing direction according to the thickness at the time of deformation and the included angle.

5. The method according to claim 4, wherein The determining the upper crystal edge line and the lower crystal edge line according to the pixel values of the pixels in the target crystal image includes: Determine the straight line where the lowermost edge covered by the second calibration frame in the second target image is located as the lower crystal edge line, and the lowermost edge of the target crystal image coincides with the lowermost edge of the second calibration frame; For each column of pixels in the target crystal image, calculate the pixel value differences between two adjacent pixels above and below successively to obtain the fifth pixel value difference until the fifth pixel value difference in each column first exceeds the preset difference range, and then stop the calculation of this column; Determine the upper edge pixels belonging to the upper edge of the molybdenum stage as the pixels corresponding to the adjacent pixels above and below where the fifth pixel value difference first exceeds the preset difference range in each column and is closer to the pixel located at the center in the target crystal image; Determine the straight line where the upper edge pixels with the most identical row coordinates are located as the upper crystal edge line.

6. The method according to claim 1, characterized in that In the first second image of each growth cycle, calibrate the second calibration frame and the position of the second crystal center, and adjust the image visual attributes of all the second images in the corresponding growth cycle according to the second calibration frame and the position of the second crystal center to obtain a second target image, including: In the first second image of each growth cycle, with the goal of covering at least all the crystals in the second viewing direction, construct a second calibration frame, and calibrate the center position of any crystal within the second calibration frame in the second viewing direction as the position of the second crystal center; Obtain the three-primary color information of the pixels at the position of the second crystal center corresponding to each growth cycle during the crystal growth process; According to the values of each primary color represented in the three-primary color information of the pixels at the position of the second crystal center and a preset expansion value, determine the extreme value of each primary color, where adding the preset expansion value to the value of each primary color gives the upper limit value of the extreme value of the primary color, and subtracting the preset expansion value from the value of each primary color gives the lower limit value of the extreme value of the primary color; According to the upper limit value and the lower limit value of each primary color, determine the primary color value range of each primary color; According to the second calibration frame and the primary color value range of each primary color, determine the parameter value of the image visual attribute corresponding to the second image to obtain the target parameter value of the image visual attribute, and adjust the image visual attributes of all the second images in the corresponding growth cycle according to the target parameter value to obtain a second target image.

7. The method according to claim 6, wherein The determining the parameter value of the image visual attribute corresponding to the second image according to the second calibration frame and the primary color value range of each primary color includes: Starting from the preset initial value of the image visual attribute, adjust the parameter value of the image visual attribute step by step according to a preset step size; After each adjustment of the parameter value, count the number of pixels within the range covered by the second calibration frame whose three primary colors are within the primary color value range of each primary color; Determine the parameter value corresponding to the largest number as the target parameter value of the image visual attribute.

8. A crystal growth apparatus, characterized in that, The crystal growth apparatus includes: a control terminal, a vacuum reaction chamber, a microwave energy supply unit, a process gas supply unit, an electric proportional valve, a vacuum pump, a first camera, and a second camera; The vacuum reaction chamber is used to provide a crystal growth environment. A molybdenum stage is arranged in the vacuum reaction chamber, and the molybdenum stage is used to carry the grown crystal. A first observation window and a second observation window are arranged in different viewing directions of the vacuum reaction chamber. The first camera is arranged at the first observation window, and the second camera is arranged at the second observation window; The first camera takes a first image of the inside of the vacuum reaction chamber from a first viewing direction through the first observation window, and the second camera takes a second image of the inside of the vacuum reaction chamber from a second viewing direction through the second observation window; The microwave power supply unit and the process gas supply unit are respectively connected to the vacuum reaction chamber through pipelines. The vacuum pump is connected to the vacuum reaction chamber through a pipeline, and an electric proportional valve is provided on the pipeline connecting the vacuum pump and the vacuum reaction chamber. The control terminal is respectively communicatively connected to the vacuum reaction chamber, the microwave energy supply unit, the process gas supply unit, the electric proportional valve, the vacuum pump, the first camera and the second camera, and is configured to control the growth of the crystal, obtain the first image and the second image, and execute the method according to any one of claims 1-7.

9. An electronic device, characterized in that, Comprising: A memory storing a computer program thereon; A processor configured to execute the computer program in the memory to implement the steps of the method according to any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the steps of the method according to any one of claims 1-7.

Images