Detection method and detection device for seeding welding image

Abstract

The invention provides a detection method and a detection device for a seeding welding image, and the detection method comprises the following steps: obtaining a target image when seeding welding is carried out on a seed crystal and a melt in a seeding process of representing that the seed crystal is utilized to guide the melt to form a single crystal material in a single crystal furnace; inputting the target image into a pre-trained image detection model, and performing aperture detection on the target image by using the image detection model to obtain an aperture detection result corresponding to the target image output by the image detection model; and inputting the aperture detection result into a preset seeding control system to obtain corresponding state information output by the seeding control system when the seed crystal and the melt are subjected to seeding welding, so as to monitor the seeding process. By means of the method, low-labor-cost monitoring of the seeding process is achieved, the accuracy of detecting the welding state in the seeding welding image is improved, and then the production efficiency and seeding quality of the seeding technology are improved.

Description

Technical Field

[0001] The present application relates to the field of seeding technology, and in particular to a method and device for detecting seeding fusion images. Background Art

[0002] At present, the existing seeding technology mainly monitors the aperture state during the welding of seed crystal and melt through manual observation. This method increases the manpower input cost, is very inefficient in the use of human resources, increases the time required for welding in the seeding technology, and reduces the production efficiency of the seeding technology.

[0003] In addition, since different operators have different judgment standards for the aperture state, the same operator's judgment results on the aperture state will fluctuate at different times and under different working conditions, which reduces the accuracy of detecting the welding state in the seeding welding image and thus reduces the seeding quality. Summary of the Invention

[0004] In view of this, the purpose of the present application is to provide a method and device for detecting seeding fusion images, by inputting the target image of the seed crystal and the melt in the single crystal furnace collected during seeding fusion into the image detection model, and using the image detection model to perform aperture detection on the target image to obtain the aperture detection result corresponding to the target image, and using the seeding control system to determine the state information of the seed crystal and the melt during seeding fusion based on the aperture detection result, so as to realize low-labor-cost monitoring of the seeding process, improve the accuracy of detecting the fusion state in the seeding fusion image, and thereby improve the production efficiency and seeding quality of the seeding technology.

[0005] The present application provides a method for detecting a seeding fusion image, the method comprising:

[0006] Characterizing a process in which a seed crystal is used to guide a melt to form a single crystal material in a single crystal furnace, obtaining a target image of the seed crystal and the melt during seeding and welding;

[0007] Inputting a target image into a pre-trained image detection model, performing aperture detection on the target image using the image detection model, and obtaining an aperture detection result corresponding to the target image output by the image detection model;

[0008] The aperture detection result is input into a preset seeding control system to obtain the corresponding state information of the seed crystal and the melt during seeding welding output by the seeding control system, so as to monitor the seeding process.

[0009] Furthermore, the detection method further comprises:

[0010] Based on the aperture detection result and the seeding parameters configured in the single crystal furnace, the seeding parameters are adjusted using the seeding control system to determine target seeding parameters, and the seeding parameters are replaced with the target seeding parameters.

[0011] Furthermore, the image detection model is pre-trained by the following steps:

[0012] performing image processing on a plurality of preset training images representing seeding and welding of the seed crystal and the melt, respectively, to obtain target training images corresponding to each of the training images;

[0013] Labeling each target training image according to the aperture display content in each target training image to obtain a target training label image corresponding to each target training image;

[0014] Inputting each of the target training label images into the to-be-trained image detection model, respectively, to obtain a first detection result corresponding to each of the target training label images output by the to-be-trained image detection model;

[0015] The model parameters in the image detection model to be trained are updated using the first detection result, so that the image detection model to be trained learns the aperture display content in the target training label image according to the preset training parameters, and obtains the image detection model corresponding to the training of the image detection model to be trained.

[0016] Furthermore, the step of pre-training the image detection model further includes:

[0017] Inputting a plurality of preset verification images characterizing the seed crystal and the melt undergoing seeding and fusion bonding into the image detection model, and obtaining a second detection result corresponding to each of the verification images output by the image detection model to be trained;

[0018] Based on the second detection result, evaluating the detection performance of the image detection model to obtain an evaluation result of the image detection model;

[0019] The evaluation results are used to optimize the model parameters in the image detection model to obtain the optimized image detection model.

[0020] Furthermore, the aperture detection result at least includes the number of apertures being a first preset value, the number of apertures being a second preset value, and the number of apertures being a third preset value.

[0021] Furthermore, the aperture detection result is input into a preset seeding control system to obtain state information corresponding to the seed crystal and the melt being seeded and welded, output by the seeding control system, including:

[0022] When the aperture detection result shows that the number of apertures is a first preset value, the aperture detection result is input into the seeding control system, and the state information output by the seeding control system is obtained as an un-fused state;

[0023] When the aperture detection result shows that the number of apertures is a second preset value, the aperture detection result is input into the seeding control system, and the state information output by the seeding control system is a normal welding state;

[0024] When the aperture detection result shows that the number of apertures is a third preset value, the aperture detection result is input into the seeding control system, and the status information output by the seeding control system is obtained as an abnormal welding state.

[0025] The present application also provides a device for detecting a seeding fusion image, the device comprising:

[0026] An image acquisition module is used to characterize the seeding process of using a seed crystal to guide the melt to form a single crystal material in a single crystal furnace, and to obtain a target image when the seed crystal and the melt are seeded and fused;

[0027] An image detection module is configured to input a target image into a pre-trained image detection model, perform aperture detection on the target image using the image detection model, and obtain an aperture detection result corresponding to the target image output by the image detection model;

[0028] A state detection module is used to input the aperture detection result into a preset seeding control system, and obtain the state information corresponding to the seed crystal and the melt when seeding and welding, which is output by the seeding control system, to monitor the seeding process.

[0029] Furthermore, the detection device further includes a parameter adjustment module, which is used to:

[0030] Based on the aperture detection result and the seeding parameters configured in the single crystal furnace, the seeding parameters are adjusted using the seeding control system to determine target seeding parameters, and the seeding parameters are replaced with the target seeding parameters.

[0031] Furthermore, when the image detection module is used to pre-train the image detection model, the image detection module is used to:

[0032] performing image processing on a plurality of preset training images representing seeding and welding of the seed crystal and the melt, respectively, to obtain target training images corresponding to each of the training images;

[0033] Labeling each target training image according to the aperture display content in each target training image to obtain a target training label image corresponding to each target training image;

[0034] Inputting each of the target training label images into the to-be-trained image detection model, respectively, to obtain a first detection result corresponding to each of the target training label images output by the to-be-trained image detection model;

[0035] The model parameters in the image detection model to be trained are updated using the first detection result, so that the image detection model to be trained learns the aperture display content in the target training label image according to the preset training parameters, and obtains the image detection model corresponding to the training of the image detection model to be trained.

[0036] Furthermore, the image detection module is also used to:

[0037] Inputting a plurality of preset verification images characterizing the seed crystal and the melt undergoing seeding and fusion bonding into the image detection model, and obtaining a second detection result corresponding to each of the verification images output by the image detection model to be trained;

[0038] Based on the second detection result, evaluating the detection performance of the image detection model to obtain an evaluation result of the image detection model;

[0039] The evaluation results are used to optimize the model parameters in the image detection model to obtain the optimized image detection model.

[0040] Furthermore, when the state detection module is used to input the aperture detection result into a preset seeding control system and obtain the state information corresponding to the seed crystal and the melt being seeded and welded, output by the seeding control system, the state detection module is used to:

[0041] When the aperture detection result shows that the number of apertures is a first preset value, the aperture detection result is input into the seeding control system, and the state information output by the seeding control system is obtained as an un-fused state;

[0042] When the aperture detection result shows that the number of apertures is a second preset value, the aperture detection result is input into the seeding control system, and the state information output by the seeding control system is a normal welding state;

[0043] When the aperture detection result shows that the number of apertures is a third preset value, the aperture detection result is input into the seeding control system, and the status information output by the seeding control system is obtained as an abnormal welding state.

[0044] An embodiment of the present application also provides an electronic device, comprising: a processor, a memory, and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor and the memory communicate through the bus, and when the machine-readable instructions are executed by the processor, the steps of the above-mentioned method for detecting seed crystal fusion images are performed.

[0045] An embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the above-mentioned method for detecting seeding fusion images are executed.

[0046] The embodiments of the present application provide a method and device for detecting seeding welding images, and the detection method includes: characterizing the seeding process in which a seed crystal is used to guide a melt to form a single crystal material in a single crystal furnace, and obtaining a target image when the seed crystal and the melt are seeded and welded; inputting the target image into a pre-trained image detection model, and performing aperture detection on the target image using the image detection model to obtain an aperture detection result corresponding to the target image output by the image detection model; inputting the aperture detection result into a preset seeding control system, and obtaining status information corresponding to the seed crystal and the melt when they are seeded and welded, output by the seeding control system, so as to monitor the seeding process.

[0047] Compared with the method of monitoring the aperture state of the seed crystal and the melt during seeding and welding by the existing seeding technology in the prior art mainly through manual observation, the target image of the seed crystal and the melt during seeding and welding in the single crystal furnace collected is input into the image detection model, and the image detection model is used to perform aperture detection on the target image to obtain the aperture detection result corresponding to the target image, and the seeding control system is used to determine the state information of the seed crystal and the melt during seeding and welding based on the aperture detection result, so as to realize low-labor-cost monitoring of the seeding process, improve the accuracy of detecting the welding state in the seeding welding image, and thereby improve the production efficiency and seeding quality of the seeding technology.

[0048] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without creative work.

[0050] Figure 1 This is one of the flow charts of a method for detecting a seeding fusion image provided in an embodiment of the present application;

[0051] Figure 2 A schematic diagram of aperture detection results of a target image provided by an embodiment of the present application;

[0052] Figure 3 A schematic diagram of annotating a target training label image provided in an embodiment of the present application;

[0053] Figure 4 This is a second flow chart of a method for detecting a seeding fusion image provided in an embodiment of the present application;

[0054] Figure 5 This is one of the structural schematic diagrams of a device for detecting seeding fusion images provided in an embodiment of the present application;

[0055] Figure 6 This is a second structural diagram of a device for detecting seeding fusion images provided in an embodiment of the present application;

[0056] Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for which protection is claimed, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, each other embodiment obtained by those skilled in the art without making creative work falls within the scope of protection of the present application.

[0058] Research has found that the current method of monitoring the aperture state during the welding of seed crystal and melt in existing seeding technology is mainly through manual observation, that is, a dedicated person is required to continuously observe the state of the aperture produced during welding. This method increases the cost of manpower input and is very inefficient in the use of human resources. Manual observation prolongs the waiting time for judgment during the seeding process, increases the time required for welding in the seeding technology, and reduces the production efficiency of the seeding technology.

[0059] In addition, since different operators have different judgment standards for the aperture state, the same operator's judgment results on the aperture state will fluctuate at different times and under different working conditions, which reduces the accuracy of detecting the welding state in the seeding welding image and thus reduces the seeding quality.

[0060] Based on this, an embodiment of the present application provides a method for detecting a seeding fusion image, by inputting the target image of the seed crystal and the melt in the single crystal furnace collected during seeding fusion into an image detection model, and using the image detection model to perform aperture detection on the target image to obtain the aperture detection result corresponding to the target image, and using the seeding control system to determine the state information of the seed crystal and the melt during seeding fusion based on the aperture detection result, so as to realize low-labor-cost monitoring of the seeding process, improve the accuracy of detecting the fusion state in the seeding fusion image, and thereby improve the production efficiency and seeding quality of the seeding technology.

[0061] See also Figure 1 , Figure 1 This is one of the flow charts of a method for detecting a seeding fusion image provided in an embodiment of the present application. Figure 1 As shown in , the method for detecting seeding fusion images provided in the embodiment of the present application includes:

[0062] S101 , characterizing a seeding process in a single crystal furnace in which a seed crystal is used to guide a melt to form a single crystal material, and obtaining a target image of the seed crystal and the melt during seeding and welding.

[0063] It should be noted that seeding technology is a key technology in the field of semiconductors and crystal growth, mainly used to manufacture high-quality single crystal materials. The seeding process usually involves immersing a seed crystal (a small piece of single crystal) with the same composition into the molten raw material, and then guiding the melt to solidify in an orderly manner according to the lattice structure of the seed crystal through slow cooling or other methods, thereby forming a larger single crystal.

[0064] Among them, the seed crystal is the "seed" of single crystal growth. It is a small crystal with a specific crystal orientation. It provides a template for crystal growth, ensuring that the atomic arrangement of the new crystal is consistent with the seed crystal, thereby forming high-quality single crystal materials. After the seed crystal is preheated and contacts the melt, the atoms in the melt are arranged according to the lattice structure of the seed crystal, and a single crystal is gradually grown; the melt refers to a uniform liquid substance formed by the melting of a substance at high temperature. It is a key medium in crystal growth, metallurgy, material processing and other processes.

[0065] A single crystal furnace is used to provide a high-temperature environment to melt raw materials (e.g., silicon and sapphire) and control the crystal growth process. The components of a single crystal furnace may include a crucible, a heating system, graphite parts, and a quartz screen.

[0066] Here, the seed crystal and the melt are seeded at the seeding melting point. The seeding melting point is the area where the seed crystal and the melt first come into contact and form a stable solid-liquid interface, which is used to provide a lattice template for crystal growth. By controlling the temperature and interface shape (aperture, a meniscus) of the seeding melting point, the uniformity of subsequent single crystal growth is ensured.

[0067] Among them, the seeding melting point is generally located on the surface of the melt in the crucible of the single crystal furnace, that is, the area where the end of the seed crystal contacts the melt. During the melting stage of the seeding process, the seed crystal will drop to a certain distance from the melt surface for preheating (baking), and then contact the melt to form a solid-liquid interface, forming a stable meniscus aperture (seeding melting point).

[0068] In this step, during the seeding process in the single crystal furnace, a camera device located near the crucible in the single crystal furnace is used to capture images of the seed crystal and the melt during seeding and melting, and the images are screened to obtain a target image, ensuring that the target image can clearly capture the aperture state of the seed crystal and the melt during seeding and melting.

[0069] S102: Input the target image into a pre-trained image detection model, use the image detection model to perform aperture detection on the target image, and obtain an aperture detection result corresponding to the target image output by the image detection model.

[0070] In an embodiment of the present application, the image detection model may include a target detection model, for example, a YOLOv5 model. The YOLOv5 model may include a backbone network (Backbone), a feature fusion (Neck) and a detection head (Head); wherein the backbone network part is used to extract basic feature information of the image, the feature fusion part is used to enhance the features extracted by the backbone network part, and the detection head part is used to perform target detection based on the feature map output by the feature fusion part, and output the position, category and confidence score of the prediction box.

[0071] Here, the aperture detection result at least includes the number of apertures being a first preset value, the number of apertures being a second preset value, and the number of apertures being a third preset value.

[0072] Here, the first preset value is generally set to 0, the second preset value is generally set to 1, and the third preset value is generally set to 2. It can also be set to other values according to the image display situation and detection requirements. This application will not limit it here.

[0073] For example, see Figure 2 , Figure 2 This is a schematic diagram of the aperture detection result of a target image provided by an embodiment of the present application. Figure 2As shown in , the aperture detection result of target image A is that the number of apertures is the second preset value (1), and the confidence score is 0.90; the aperture detection result of target image B is that the number of apertures is the first preset value (0), and the confidence score is 0.92; the aperture detection result of target image C is that the number of apertures is the third preset value (2), and the confidence score is 0.94.

[0074] In one embodiment of the present application, in specific implementation, the step of pre-training the image detection model in step S102 may include:

[0075] S102A: performing image processing on a plurality of preset training images representing seeding and welding of the seed crystal and the melt, respectively, to obtain a target training image corresponding to each of the training images.

[0076] In this step, image processing is performed on a plurality of preset training images respectively to obtain a target training image corresponding to each training image after image processing.

[0077] Among them, each training image is a pre-collected image of the seed crystal and the melt undergoing seeding and welding.

[0078] In an embodiment of the present application, the image processing method may include but is not limited to image data cleaning (to remove irrelevant or low-quality data), image resizing (to unify the image size and aspect ratio), normalization (to standardize the mean and standard deviation) and image data enhancement (to ensure the diversity of training images), etc.

[0079] S102B: label each target training image according to the aperture display content in each target training image to obtain a target training label image corresponding to each target training image.

[0080] In an embodiment of the present application, the aperture display content in each target training image may include but is not limited to the number of apertures being a first preset value, the number of apertures being a second preset value, and the number of apertures being a third preset value.

[0081] Here, the first preset value is generally set to 0, the second preset value is generally set to 1, and the third preset value is generally set to 2.

[0082] In this step, the aperture display content in each target training image is annotated, that is, the number of apertures in each target training image is marked to obtain the target training label image corresponding to each annotated target training image.

[0083] For example, see Figure 3 , Figure 3 This is a diagram of a target training label image provided by an embodiment of the present application. Figure 3As shown in , the aperture display content annotated by the target training label image B5020 is that the number of apertures is 0, the aperture display content annotated by the target training label image B5027 is that the number of apertures is 1, and the aperture display content annotated by the target training label image B5106 is that the number of apertures is 2.

[0084] S102C: Input each of the target training label images into the to-be-trained image detection model, and obtain a first detection result corresponding to each of the target training label images output by the to-be-trained image detection model.

[0085] In this step, each target training label image is input into the image detection model to be trained respectively. After the image detection model to be trained detects and classifies each target training label image, the first detection result corresponding to each target training label image output by the image detection model to be trained is obtained.

[0086] Among them, the first detection result corresponding to each target training label image corresponds to the aperture display content in each target training image, and the first detection result corresponding to each target training label image may include but is not limited to the number of apertures being a first preset value, the number of apertures being a second preset value, and the number of apertures being a third preset value.

[0087] Here, the first preset value is generally set to 0, the second preset value is generally set to 1, and the third preset value is generally set to 2.

[0088] S102D. Use the first detection result to update the model parameters in the image detection model to be trained, so that the image detection model to be trained learns the aperture display content in the target training label image according to the preset training parameters, and obtains the image detection model corresponding to the training of the image detection model to be trained.

[0089] In this step, the accuracy of the first detection result corresponding to each target training label image is calibrated for the image detection model to be trained, and the model parameters in the image detection model to be trained are updated so that the image detection model to be trained learns the aperture display content in the target training label image according to the preset training parameters, and then updates the model parameters in the image detection model to be trained to obtain the image detection model corresponding to the image detection model to be trained.

[0090] In an embodiment of the present application, the training parameters include at least input image size, batch number of images, learning rate and training cycle.

[0091] Optionally, in addition to the steps of pre-training the image detection model described in steps S102A to S102D, the embodiment of the present application also includes steps S102E to S102G. Specifically, steps S102E to S102G are used to illustrate a method for verifying and optimizing the image detection model to improve the detection performance of the image detection model.

[0092] S102E. Input a plurality of preset verification images representing the seed crystal and the melt undergoing seeding and fusion bonding into the image detection model to obtain a second detection result corresponding to each verification image output by the image detection model to be trained.

[0093] In this step, after the image detection model to be trained is trained according to the preset training parameters to obtain the image detection model, the image detection model is evaluated and verified to test the detection performance of the image detection model, that is, a preset plurality of verification images of seed crystals and melts for seeding and welding are input into the image detection model, and the image detection model detects and classifies each verification image to obtain a second detection result corresponding to each verification image output by the image detection model.

[0094] Among them, each verification image is a pre-collected image of the seed crystal and the melt undergoing seeding and welding.

[0095] Here, the second detection result corresponding to each verification image corresponds to the first detection result corresponding to each target training label image, and the first detection result corresponding to each target training label image may include but is not limited to the number of apertures being the first preset value, the number of apertures being the second preset value, and the number of apertures being the third preset value.

[0096] Here, the first preset value is generally set to 0, the second preset value is generally set to 1, and the third preset value is generally set to 2.

[0097] S102F. Based on the second detection result, evaluate the detection performance of the image detection model to obtain an evaluation result of the image detection model.

[0098] In this step, based on the second detection result corresponding to each verification image and the actual classification result corresponding to each verification image, an evaluation result representing the detection performance of the evaluation image detection model is obtained.

[0099] In an embodiment of the present application, the evaluation result of the image detection model may include a detection accuracy rate of the image detection model.

[0100] S102G. Optimize the model parameters in the image detection model using the evaluation result to obtain the optimized image detection model.

[0101] In this step, according to the excellent evaluation results, the model parameters in the image detection model are adjusted to optimize the detection performance of the image detection model and obtain the optimized image detection model.

[0102] S103, inputting the aperture detection result into a preset seeding control system, obtaining the state information corresponding to the seed crystal and the melt during seeding welding output by the seeding control system, so as to monitor the seeding process.

[0103] In the embodiment of the present application, the aperture in the aperture detection result corresponding to the target image is an optical phenomenon in the contact area between the seed crystal and the melt during the single crystal growth process. When the end of the seed crystal contacts the melt surface and forms a stable solid-liquid interface, a visible circular or annular aperture will be formed due to the temperature gradient at the interface, melt flow and light refraction effect.

[0104] Here, the aperture detection result corresponding to the target image can reflect the corresponding state information when the seed crystal and the melt are seeded and welded.

[0105] Among them, the seeding control system refers to a system used to precisely control the contact process between the seed crystal and the melt in the single crystal growth process. Its core functions include: temperature control, pulling speed and rotation speed adjustment, aperture monitoring and automation optimization, etc.

[0106] In one embodiment of the present application, during specific implementation, step S103 may include:

[0107] S1031. When the aperture detection result shows that the number of apertures is a first preset value, the aperture detection result is input into the seeding control system, and the state information output by the seeding control system is obtained as an un-fused state.

[0108] In an embodiment of the present application, when the aperture detection result shows that the number of apertures is the first preset value, that is, when the aperture detection result shows that the number of apertures is 0, it indicates that the corresponding state information when the seed crystal and the melt are subjected to seeding and melting is an unmelted state, that is, a stable solid-liquid interface is not formed between the seed crystal and the melt, or the formed solid-liquid interface is in an unstable state.

[0109] Here, the reasons for the state information corresponding to the unfused state when the seed crystal and the melt are fused during seeding may include but are not limited to fusion temperature mismatch, excessive thermal shock, and melt convection interference.

[0110] S1032. When the aperture detection result shows that the number of apertures is a second preset value, the aperture detection result is input into the seeding control system, and the status information output by the seeding control system is obtained as a normal welding state.

[0111] In an embodiment of the present application, when the aperture detection result shows that the number of apertures is the second preset value, that is, when the aperture detection result shows that the number of apertures is 1, it indicates that the corresponding state information when the seed crystal and the melt are subjected to seeding and melting is a normal melting state, that is, the seed crystal and the melt form a stable solid-liquid interface, and the interface appears as a single bright halo.

[0112] Here, the reasons for the state information corresponding to the normal welding state when the seed crystal and the melt are welded during seeding may include but are not limited to successful welding and controllable crystal diameter.

[0113] S1033. When the aperture detection result shows that the number of apertures is a third preset value, the aperture detection result is input into the seeding control system, and the status information output by the seeding control system is obtained as an abnormal welding state.

[0114] In an embodiment of the present application, when the aperture detection result shows that the number of apertures is the third preset value, that is, when the aperture detection result shows that the number of apertures is 2, it indicates that the corresponding state information when the seed crystal and the melt are subjected to seeding and melting is an abnormal melting state, that is, two separate light rings appear at the interface, causing the interface to become unstable or the melt flow to be abnormal.

[0115] Here, the reasons for the abnormal welding state corresponding to the state information of the seed crystal and the melt during seeding welding may include but are not limited to excessive melt convection, uneven thermal stress, and interference from impurities or bubbles.

[0116] Optional, see Figure 4 , Figure 4 This is a second flow chart of a method for detecting a seeding fusion image provided in an embodiment of the present application. Figure 4 As shown in , the method for detecting the seeding fusion image provided in the embodiment of the present application further includes step S104 in addition to steps S101 to S103. Specifically, step S104 is used to illustrate a method for optimizing the seeding parameters of the seeding control system based on the aperture detection results to improve the performance of the seeding control system.

[0117] S104. Based on the aperture detection result and the seeding parameters configured in the single crystal furnace, adjust the seeding parameters using the seeding control system to determine target seeding parameters, and replace the seeding parameters with the target seeding parameters.

[0118] In the embodiment of the present application, the adjusted seeding parameters include at least temperature parameters, pulling speed and rotation speed parameters, crucible position and rotation parameters, etc.

[0119] In this step, when adjusting the temperature parameter in the seeding parameter, when the aperture detection result shows that the number of apertures is a first preset value, the temperature of the melt is increased; when the aperture detection result shows that the number of apertures is a third preset value, the temperature of the melt is decreased.

[0120] Furthermore, when adjusting the pulling speed and rotation speed parameters in the seeding parameters, when the aperture detection result shows that the number of apertures is a first preset value, the pulling speed is increased; when the aperture detection result shows that the number of apertures is a third preset value, the pulling speed is reduced.

[0121] Furthermore, when adjusting the crucible position and rotation parameters in the seeding parameters, when the aperture detection result shows that the number of apertures is a first preset value, the crucible position is adjusted; when the aperture detection result shows that the number of apertures is a third preset value, the crucible rotation speed is reduced.

[0122] The method for detecting the seeding fusion image provided in the embodiment of the present application inputs the target image of the seed crystal and the melt in the single crystal furnace collected into the image detection model, uses the image detection model to perform aperture detection on the target image, obtains the aperture detection result corresponding to the target image, and uses the seeding control system to determine the state information of the seed crystal and the melt when they are seeded and fused based on the aperture detection result, so as to realize low-manpower cost monitoring of the seeding process, improve the accuracy of detecting the fusion state in the seeding fusion image, and thereby improve the production efficiency and seeding quality of the seeding technology.

[0123] See also Figure 5 、 Figure 6 , Figure 5 This is one of the structural schematic diagrams of a device for detecting a seeding fusion image provided in an embodiment of the present application. Figure 6 This is a second structural diagram of a device for detecting seeding fusion images provided in an embodiment of the present application. Figure 5 As shown in , the detection device 500 includes:

[0124] An image acquisition module 510 is used to characterize a seeding process in which a seed crystal is used to guide a melt to form a single crystal material in a single crystal furnace, and to obtain a target image of the seed crystal and the melt during seeding and fusion bonding;

[0125] An image detection module 520 is configured to input a target image into a pre-trained image detection model, perform aperture detection on the target image using the image detection model, and obtain an aperture detection result corresponding to the target image output by the image detection model;

[0126] The state detection module 530 is used to input the aperture detection result into a preset seeding control system, and obtain the state information corresponding to the seed crystal and the melt during seeding welding output by the seeding control system to monitor the seeding process.

[0127] Further, such as Figure 6 As shown in , the detection device 500 further includes a parameter adjustment module 540, and the parameter adjustment module 540 is used to:

[0128] Based on the aperture detection result and the seeding parameters configured in the single crystal furnace, the seeding parameters are adjusted using the seeding control system to determine target seeding parameters, and the seeding parameters are replaced with the target seeding parameters.

[0129] Furthermore, when the image detection module 520 is used to pre-train the image detection model, the image detection module 520 is used to:

[0130] performing image processing on a plurality of preset training images representing seeding and welding of the seed crystal and the melt, respectively, to obtain target training images corresponding to each of the training images;

[0131] Labeling each target training image according to the aperture display content in each target training image to obtain a target training label image corresponding to each target training image;

[0132] Inputting each of the target training label images into the to-be-trained image detection model, respectively, to obtain a first detection result corresponding to each of the target training label images output by the to-be-trained image detection model;

[0133] The model parameters in the image detection model to be trained are updated using the first detection result, so that the image detection model to be trained learns the aperture display content in the target training label image according to the preset training parameters, and obtains the image detection model corresponding to the training of the image detection model to be trained.

[0134] Furthermore, the image detection module 520 is further configured to:

[0135] Inputting a plurality of preset verification images characterizing the seed crystal and the melt undergoing seeding and fusion bonding into the image detection model, and obtaining a second detection result corresponding to each of the verification images output by the image detection model to be trained;

[0136] Based on the second detection result, evaluating the detection performance of the image detection model to obtain an evaluation result of the image detection model;

[0137] The evaluation results are used to optimize the model parameters in the image detection model to obtain the optimized image detection model.

[0138] Furthermore, when the state detection module 530 is used to input the aperture detection result into a preset seeding control system and obtain the state information corresponding to the seed crystal and the melt being seeded and welded, output by the seeding control system, the state detection module 530 is used to:

[0139] When the aperture detection result shows that the number of apertures is a first preset value, the aperture detection result is input into the seeding control system, and the state information output by the seeding control system is obtained as an un-fused state;

[0140] When the aperture detection result shows that the number of apertures is a second preset value, the aperture detection result is input into the seeding control system, and the state information output by the seeding control system is a normal welding state;

[0141] When the aperture detection result shows that the number of apertures is a third preset value, the aperture detection result is input into the seeding control system, and the status information output by the seeding control system is obtained as an abnormal welding state.

[0142] The detection device for the seeding fusion image provided in the embodiment of the present application inputs the target image of the seed crystal and the melt in the single crystal furnace collected into the image detection model, uses the image detection model to perform aperture detection on the target image, obtains the aperture detection result corresponding to the target image, and uses the seeding control system to determine the state information of the seed crystal and the melt when they are seeded and fused based on the aperture detection result, so as to realize low-manpower cost monitoring of the seeding process, improve the accuracy of detecting the fusion state in the seeding fusion image, and thus improve the production efficiency and seeding quality of the seeding technology.

[0143] See also Figure 7 , Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 7 As shown in FIG, the electronic device 700 includes a processor 710 , a memory 720 and a bus 730 .

[0144] The memory 720 stores machine-readable instructions executable by the processor 710. When the electronic device 700 is running, the processor 710 communicates with the memory 720 via the bus 730. When the machine-readable instructions are executed by the processor 710, the above-mentioned Figure 1 as well as Figure 4 The specific implementation of the steps of the method for detecting the seeding fusion image in the method embodiment shown can be found in the method embodiment, and will not be repeated here.

[0145] The embodiment of the present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the computer program can execute the above-mentioned Figure 1 as well as Figure 4 The specific implementation of the steps of the method for detecting the seeding fusion image in the method embodiment shown can be found in the method embodiment, and will not be repeated here.

[0146] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0147] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. There may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed may be through some communication interface, indirect coupling or communication connection of devices or units, which may be electrical, mechanical or other forms.

[0148] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0149] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0150] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0151] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The scope of protection of the present application is not limited thereto. Although the present application has been described in detail with reference to the above-mentioned embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily conceive of changes to the technical solutions described in the above-mentioned embodiments within the technical scope disclosed in the present application, or perform equivalent replacements for some of the technical features thereof. These modifications, changes, or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A method for detecting a seeding fusion image, characterized in that: The detection method comprises: Characterizing a process in which a seed crystal is used to guide a melt to form a single crystal material in a single crystal furnace, obtaining a target image of the seed crystal and the melt during seeding and welding; Inputting a target image into a pre-trained image detection model, performing aperture detection on the target image using the image detection model, and obtaining an aperture detection result corresponding to the target image output by the image detection model; The aperture detection result is input into a preset seeding control system to obtain the corresponding state information of the seed crystal and the melt during seeding welding output by the seeding control system, so as to monitor the seeding process.

2. The method according to claim 1, characterized in that The detection method further comprises: Based on the aperture detection result and the seeding parameters configured in the single crystal furnace, the seeding parameters are adjusted using the seeding control system to determine target seeding parameters, and the seeding parameters are replaced with the target seeding parameters.

3. The method according to claim 1, characterized in that The image detection model is pre-trained by the following steps: performing image processing on a plurality of preset training images representing seeding and welding of the seed crystal and the melt, respectively, to obtain target training images corresponding to each of the training images; Labeling each target training image according to the aperture display content in each target training image to obtain a target training label image corresponding to each target training image; Inputting each of the target training label images into the to-be-trained image detection model, respectively, to obtain a first detection result corresponding to each of the target training label images output by the to-be-trained image detection model; The model parameters in the image detection model to be trained are updated using the first detection result, so that the image detection model to be trained learns the aperture display content in the target training label image according to the preset training parameters, and obtains the image detection model corresponding to the training of the image detection model to be trained.

4. The method according to claim 3, characterized in that The step of pre-training the image detection model further includes: Inputting a plurality of preset verification images characterizing the seed crystal and the melt undergoing seeding and fusion bonding into the image detection model, and obtaining a second detection result corresponding to each of the verification images output by the image detection model to be trained; Based on the second detection result, evaluating the detection performance of the image detection model to obtain an evaluation result of the image detection model; The evaluation results are used to optimize the model parameters in the image detection model to obtain the optimized image detection model.

5. The method according to claim 1, wherein The aperture detection result at least includes the number of apertures being a first preset value, the number of apertures being a second preset value, and the number of apertures being a third preset value.

6. The method according to claim 5, characterized in that The step of inputting the aperture detection result into a preset seeding control system to obtain state information corresponding to the seed crystal and the melt being seeded and welded, output by the seeding control system, includes: When the aperture detection result shows that the number of apertures is a first preset value, the aperture detection result is input into the seeding control system, and the state information output by the seeding control system is obtained as an un-fused state; When the aperture detection result shows that the number of apertures is a second preset value, the aperture detection result is input into the seeding control system, and the state information output by the seeding control system is a normal welding state; When the aperture detection result shows that the number of apertures is a third preset value, the aperture detection result is input into the seeding control system, and the status information output by the seeding control system is obtained as an abnormal welding state.

7. A device for detecting seeding fusion images, characterized in that: The detection device comprises: An image acquisition module is used to characterize the seeding process of using a seed crystal to guide the melt to form a single crystal material in a single crystal furnace, and to obtain a target image when the seed crystal and the melt are seeded and fused; An image detection module is configured to input a target image into a pre-trained image detection model, perform aperture detection on the target image using the image detection model, and obtain an aperture detection result corresponding to the target image output by the image detection model; A state detection module is used to input the aperture detection result into a preset seeding control system, and obtain the state information corresponding to the seed crystal and the melt when seeding and welding, which is output by the seeding control system, to monitor the seeding process.

8. The device according to claim 7, characterized in that The detection device further includes a parameter adjustment module, which is configured to: Based on the aperture detection result and the seeding parameters configured in the single crystal furnace, the seeding parameters are adjusted using the seeding control system to determine target seeding parameters, and the seeding parameters are replaced with the target seeding parameters.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor and the memory communicate via the bus. When the processor runs the machine-readable instructions, the steps of the method for detecting a seeding fusion image as described in any one of claims 1 to 6 are executed.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for detecting seeding fusion images according to any one of claims 1 to 6 are executed.

Images