CN113781474A - Collapse state detection method and device and storage medium - Google Patents
Collapse state detection method and device and storage medium Download PDFInfo
- Publication number
- CN113781474A CN113781474A CN202111159140.XA CN202111159140A CN113781474A CN 113781474 A CN113781474 A CN 113781474A CN 202111159140 A CN202111159140 A CN 202111159140A CN 113781474 A CN113781474 A CN 113781474A
- Authority
- CN
- China
- Prior art keywords
- displacement
- image
- angular point
- corner
- production image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20016—Hierarchical, coarse-to-fine, multiscale or multiresolution image processing; Pyramid transform
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20112—Image segmentation details
- G06T2207/20164—Salient point detection; Corner detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30148—Semiconductor; IC; Wafer
Abstract
The application discloses a method and a device for detecting a collapsed state and a storage medium, which relate to the technical field of semiconductor manufacturing, and the method comprises the following steps: acquiring a production image of a melting stage in real time; detecting corner information of the molten material according to the production image; determining angular point displacement according to the production image and the angular point information; calculating the total angular point displacement on the production image according to the angular point displacement; and when the total angular point displacement reaches a target threshold value, determining that the molten material stage is in a collapsed state. The problem that detection cannot be performed possibly when judgment is performed according to the number of the angular points in the prior art is solved, and the effect of detecting the collapse state in real time is achieved.
Description
Technical Field
The invention relates to a method and a device for detecting a collapsed state and a storage medium, belonging to the technical field of semiconductor manufacturing.
Background
In the field of photovoltaic industry, photovoltaic cells are power generation system units made on the basis of silicon materials, monocrystalline silicon being one of the most important substrates of photovoltaic cells, which are mostly produced in a single-product furnace by the czochralski method. The production process of the monocrystalline silicon mainly comprises the steps of melting, seeding, necking, shouldering, equal-diameter growth and ending. In the whole production process, the melting process refers to a process of melting the polycrystalline silicon material in the crucible by a heater in the single crystal furnace through heating the quartz crucible after operations such as charging, leakage detection and the like are completed. The process is the process with the highest temperature in the single crystal production process, and the time required for melting the polycrystalline silicon is longer, so the process is the process with the highest accident occurrence risk in the single crystal production process, and the accidents include silicon leakage, crucible deformation, silicon spraying, more impurities generated in the material melting process and the like, and are one of the main reasons for causing higher production cost.
In the melting process, the crucible is firstly positioned at a low crucible position and a low crucible rotation speed, so that the whole silicon material can absorb heat as much as possible. The silicon material is generally melted from the bottom, and when the solution appears at the bottom of the crucible, the rotation speed of the crucible is increased, the crucible is heated uniformly, and the deformation of the crucible is avoided. Because the temperature in the solution is higher, the melting speed of the silicon material in the solution is higher, and the silicon material on the liquid surface collapses due to the loss of the support, which is called as 'collapse'. On one hand, the whole silicon material after collapse is already in a low-temperature region at the lower part of the heater, if the silicon material is in a lower-temperature part in a thermal field for a long time, the melting time is prolonged due to insufficient heat, impurities in the solution are increased, and the silicon material enters a prepared single crystal silicon rod through melt convection during crystal pulling, so that the content of oxygen and other impurity elements in a single crystal silicon finished product is increased; in addition, the reaction time of the quartz crucible and the silicon liquid at a higher temperature is prolonged due to the prolonged melting time, so that the corrosion of the quartz crucible is increased. On the other hand, after the material collapses, the upper silicon material quickly sinks to the bottom, and because the temperature of the silicon material sinking to the bottom is different from that of the silicon solution, the temperature gradient between the silicon material sinking to the bottom and the silicon solution changes violently, and the phenomenon that the silicon solution splashes is easily caused by continuously maintaining the original crucible position (original heating state). Therefore, after the material collapse phenomenon occurs in the material melting process, the crucible position should be timely lifted, so that the whole stockpile is continuously positioned in a high-temperature area, the impurities in the completely melted silicon solution are further reduced, the original temperature gradient can be broken, and accidents such as silicon liquid splashing and the like caused by the convection temperature gradient are prevented.
The conventional method for detecting the collapsed state comprises the following steps: acquiring a current moment image frame, detecting characteristic points (namely angular points) by using a Harris operator, counting the number of the angular points on the image, and determining the melting state of a silicon material corresponding to the current moment image frame according to the number of the angular points. However, in the earlier stage of melting, the bottom of the silicon material is melted quickly, and the method cannot detect the number of corner points at the lower part of the silicon material, that is, the method cannot detect the state of the melted material. Meanwhile, in the melting process, the number of the angular points is increased and then reduced, and the melting state cannot be completely detected only through the number of the angular points.
Disclosure of Invention
The invention aims to provide a method and a device for detecting a collapsed material state and a storage medium, which are used for solving the problems in the prior art.
In order to achieve the purpose, the invention provides the following technical scheme:
according to a first aspect, an embodiment of the present invention provides a method for detecting a collapsed state, where the method includes:
acquiring a production image of a melting stage in real time;
detecting corner information of the molten material according to the production image;
determining angular point displacement according to the production image and the angular point information;
calculating the total angular point displacement on the production image according to the angular point displacement;
and when the total angular point displacement reaches a target threshold value, determining that the molten material stage is in a collapsed state.
Optionally, the method further includes:
acquiring angular point displacements acquired in a melting stage;
calculating the total displacement of the production image according to the obtained angular point displacement;
and determining the target threshold according to the total displacement of the production image.
Optionally, the determining the target threshold according to the total displacement of the production image includes:
sorting the total displacement of the obtained production images according to an ascending order, and obtaining a first displacement in a first sorting and a second displacement in a second sorting;
calculating a displacement distance of the first displacement and the second displacement;
and determining to obtain the target threshold according to the second displacement and the displacement distance.
Optionally, the detecting corner information of the frit according to the production image includes:
corner information in the production image is detected by a Shi-Tomasi corner detection algorithm.
Optionally, the determining the corner displacement according to the production image and the corner information includes:
and determining to obtain the angular point displacement according to the production image and the angular point information by using a Lucas-Kanad algorithm.
Optionally, the determining, by using a Lucas-Kanad algorithm, the corner displacement according to the production image and the corner information includes:
reducing the production image for n times to generate a first pyramid image of the production image, wherein the bottom layer of the first pyramid image is the production image; n is a positive integer;
and determining the angular point displacement according to the first pyramid image, the second pyramid image corresponding to the previous frame of production image and the angular point information by using a Lucas-Kanad algorithm.
Optionally, the determining, by using a Lucas-Kanad algorithm, the corner displacement according to the first pyramid image, the second pyramid image corresponding to the previous frame of production image, and the corner information includes:
initializing an existing displacement of an n +1 th layer in the second pyramid image;
for the L-th layer of the first pyramid image, obtaining first corner displacement of the L + 1-th layer in the second pyramid image calculated by the Lucas-Kanad algorithm, and calculating the existing displacement of the L-th layer in the second pyramid image according to the first corner displacement and the existing displacement of the L + 1-th layer in the second pyramid image; calculating a second angular point displacement of the L-th layer of the first pyramid image through the Lucas-Kanad algorithm, acquiring the existing displacement of the L-th layer of the first pyramid image according to the existing displacement of the L-th layer of the second pyramid image and the second angular point displacement, and if L is larger than 1, continuing to execute the step; the initial value of L is n;
and if L is 1, determining the determined existing displacement of the 1 st layer in the first pyramid image as the corner displacement.
Optionally, the acquiring a production image of the melt stage in real time includes:
collecting a collected image of the molten material stage through an image sensor;
and preprocessing and gray level conversion are carried out on the collected image, and the preprocessed gray level image is determined as the production image.
In a second aspect, there is provided a slump condition detection apparatus comprising a memory having stored therein at least one program instruction, and a processor for implementing the method according to the first aspect by loading and executing the at least one program instruction.
In a third aspect, there is provided a computer storage medium having stored therein at least one program instruction which is loaded and executed by a processor to implement the method of the first aspect.
Obtaining a production image of a melting stage in real time; detecting corner information of the molten material according to the production image; determining angular point displacement according to the production image and the angular point information; calculating the total angular point displacement on the production image according to the angular point displacement; and when the total angular point displacement reaches a target threshold value, determining that the molten material stage is in a collapsed state. The problem that detection cannot be performed possibly when judgment is performed according to the number of the angular points in the prior art is solved, and the effect of detecting the collapse state in real time is achieved.
Meanwhile, the angular points are detected through a Shi-Tomasi angular point detection algorithm, so that the angular point detection accuracy is improved, and the subsequent collapse detection accuracy is further improved.
And calculating a target threshold according to each calculated displacement, so that whether abnormal displacement occurs can be judged according to the size relation between the total displacement and the target threshold, and further the accuracy of the collapse detection is ensured.
The foregoing description is only an overview of the technical solutions of the present invention, and in order to make the technical solutions of the present invention more clearly understood and to implement them in accordance with the contents of the description, the following detailed description is given with reference to the preferred embodiments of the present invention and the accompanying drawings.
Drawings
Fig. 1 is a flowchart of a method for detecting a collapsed state according to an embodiment of the present invention.
Detailed Description
The technical solutions of the present invention will be described clearly and completely with reference to the accompanying drawings, and it should be understood that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, and are only for convenience of description and simplicity of description, but do not indicate or imply that the device or element being referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus, should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
In the description of the present invention, it should be noted that, unless otherwise explicitly specified or limited, the terms "mounted," "connected," and "connected" are to be construed broadly, e.g., as meaning either a fixed connection, a removable connection, or an integral connection; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
In addition, the technical features involved in the different embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
Referring to fig. 1, a flowchart of a method for detecting a slump state according to an embodiment of the present application is shown, where as shown in fig. 1, the method includes:
in practical implementation, the steps include:
firstly, acquiring an acquired image of the molten material stage through an image sensor;
secondly, preprocessing and gray level conversion are carried out on the collected image, and the preprocessed gray level image is determined as the production image.
Of course, according to the actual processing requirement, the image collected by the image sensor can also be directly used as the production image.
It should be added that, in order to improve the detection accuracy, after the production image is acquired, the acquired production image may be preprocessed by an edge detection algorithm or a filtering algorithm, and then the preprocessed production image is used in a subsequent step.
102, detecting corner information of the molten material according to the production image;
a plurality of corner points may form the shape of the frit in the image. In the present application, the frit state of the silicon material is represented by corner information.
The corner point is an intersection point of two edges and is used for representing the position of the change of the directions of the two edges, so that the direction and the amplitude of the gradient map of the region can be greatly changed when the corner point is slightly moved in any direction, that is, a pixel point corresponding to the local maximum in a first derivative (namely, the gradient of the gray scale map) is the corner point. Based on the idea, a window with a fixed size slightly slides in any direction at a certain position on an image, and if gray values (on a gradient map) in the window are greatly changed, an angular point exists in an area where the window is located. The specific detection steps are as follows:
first, the amount of change in the internal pixel values when the window (small image segment) is moved in both the x and y directions is calculated.
Assuming that the center of a window is located at a point (x, y) of the gray image, the gray value of the pixel at this position is I (x, y), if the window is moved by small displacements u and v in the x and y directions to a new position (x + u, y + v), the gray value of the pixel at this position is I (x + u, y + v), and the change value of the gray value caused by the window movement is I (x + u, y + v)
|I(x+u,y+v)-I(x,y)|. (1.1)
Let w (x, y) be the window function at location point (x, y) and represent the weight of each pixel within the window. If the pixel at the center point of the window is the angular point, the gray value of the center point changes greatly before and after the window moves, the weight coefficient of the point is large, and the contribution of the point to the gray value change is large; points far away from the center (corner) of the window have smaller gray scale change, and the weight coefficient is smaller to show that the point has smaller contribution to the gray scale change; therefore, w (x, y) can be set to a gaussian distribution (binary normal distribution) with the center of the window as the origin, and the window function distribution obeys the gaussian distribution in this embodiment as an example. The variance of the pixel grey value caused by the movement of the window in various directions (u, v) can be described by the sum of the squares of the grey levels of the image in four directions:
E(u,v)＝∑(x,y)w(x,y)×[I(x+u,y+v)-I(x,y)]2 (1.2)
in the flat region, the shift gray change of the window in any direction is small, so the E value is small; in the edge area, the image change in the moving direction of the parallel edge is small, and the image is greatly changed in the translation direction perpendicular to the edge, so that the E value is large or small; at the corner points, each directional translation greatly changes the image, and thus the E value is large.
Then, the formula (1.1) is developed by using the two-dimensional taylor formula to obtain:
I(x+u,y+v)≈I(x,y)+uIx+vIy，
wherein, IxAnd IyIs the partial differential of I. Substituting the above equation into (1.2) can be calculated as:
the matrix M is:
finally, Shi-Tomasi finds that the stability of the corner is related to the smaller eigenvalues of the matrix M, and then directly uses the smaller eigenvalues as the corner scores for the pixel. Thus, the corner point score at pixel point (x, y) is formulated as:
wherein λ1,λ2Two eigenvalues of the matrix M, respectively. Therefore, the matrix M needs to be subjected to eigenvalue decomposition, and the minimum eigenvalue is calculated as the corner score, if the score is greater than the set corner threshold, it is considered as a corner, otherwise it is not. That is, the specific corner detection method for detecting the corner information in the production image by using the Shi-Tomasi corner detection algorithm is as follows:
inputting: grayscale image, window size 3, window function w (x, y), corner threshold;
and (3) outputting: and angular point fraction and angular point position information corresponding to each pixel.
103, determining angular point displacement according to the production image and the angular point information;
and determining to obtain the angular point displacement according to the production image and the angular point information by using a Lucas-Kanad algorithm.
The Lucas-Kanad (L-K) algorithm utilizes the corner points extracted by the Shi-Tomasi method to calculate the optical flow vectors at the pixel points, and the method can quickly calculate the optical flow and has small calculation amount. The specific calculation mode of the L-K algorithm is as follows:
assuming that a target corner pixel has a gray level of I (x, y, t) at time t, at time t + δ t, the pixel has the same gray level as that at time t after moving (δ x, δ y), i.e.
I(x,y,t)＝I(x+δx,y+δy,t+δt)。 (2.1)
An optical flow equation can be calculated for (2.1),
Ixu+Iyv+It＝0， (2.2)
wherein the content of the first and second substances,
and is
Further:
wherein the content of the first and second substances,
representing the speed at which the corner pixels move,
representing the corner pixel spatial gray differential.
The optical flow equation (2.2) holds for all pixels within a window centered on the point (x, y) and the pixels within the window have the same velocity (u, v). If we use an n × n window, we get n × n equations:
where n is 25, the velocity v:
in the process of melting, when the time intervals for collecting the images are equal, the angular point displacement can be judged according to the obtained angular point speed of the pixel points. Therefore, the calculation method for determining and obtaining the angular point displacement according to the production image and the angular point information by using the Lucas-Kanad algorithm comprises the following steps:
inputting: the gray-scale images of the previous frame and the current frame, the corner information on the image of the previous frame and the window size are produced. Wherein the window size is a default value.
And (3) outputting: displacement of the corner points of the part on the previous frame image.
As can be seen from the above, the L-K algorithm needs to satisfy the assumption that the moving object moves slowly, while discontinuous motion may exist between the corner points of the molten material between adjacent image frames during the collapse, and the L-K algorithm may have a large error, so in this application, the step of obtaining the corner point displacement may further include:
firstly, reducing the production image for n times to generate a first pyramid image of the production image, wherein the bottom layer of the first pyramid image is the production image; n is a positive integer;
in one embodiment, when the pixel of the current production image is 400 × 400, the object velocity is [8,8 ]; when the image pixel is reduced to 200 × 200, the speed becomes [4,4 ]; at a zoom of 100 x 100, the velocity decreases to [2,2 ]. By analogy, the pyramid image can be obtained after the production image is reduced for multiple times.
Secondly, determining the angular point displacement according to the first pyramid image, a second pyramid image corresponding to a previous frame of production image and the angular point information through a Lucas-Kanad algorithm.
Forming two image pyramids with equal layer number by the previous frame production image I and the current frame production image I (t + delta t) through a down-sampling method
And
calculating the pixel value after down sampling by using an average value or a maximum value; the resolution of the sampled or each image pyramid is gradually reduced from bottom to top, i.e. the bottom layer resolution is the highest,the number of pixels is the largest, the resolution of the top layer is the lowest, and the number of pixels is the smallest. Wherein L ismIs the number of the topmost layer, i.e. Lm＝n+1。
Specifically, the method comprises the following steps:
(1) initializing the existing displacement of the (n + 1) th layer in the second pyramid image;
let the current displacement of the previous frame image of the L-th layer be
When L is equal to LmWhen the image is processed, the highest layer existing displacement of the previous frame image is initialized,
computing the offset of the highest layer using the L-K algorithm
(2) For the L-th layer of the first pyramid image, acquiring first corner displacement of the L + 1-th layer in the second pyramid image calculated by the Lucas-Kanad algorithm, and calculating the existing displacement of the L-th layer in the second pyramid image according to the first corner displacement and the existing displacement of the L + 1-th layer in the second pyramid image; calculating a second angular point displacement of the L-th layer of the first pyramid image through the Lucas-Kanad algorithm, acquiring the existing displacement of the L-th layer of the first pyramid image according to the existing displacement of the L-th layer of the second pyramid image and the second angular point displacement, and if L is larger than 1, continuing to execute the step; the initial value of L is n;
for the L-th layer, the L-th layer existing displacement, g, of the previous frame image is calculatedL＝2(gL+1+dL+1) (ii) a Calculating the offset of the L-th layer by using L-K algorithm
Total displacement of the L-th layer, g ═ g (g)L+dL)。
(3) And if L is 1, determining the determined existing displacement of the 1 st layer in the first pyramid image as the corner displacement.
104, calculating the total angular point displacement on the production image according to the angular point displacement;
and summing the calculated angular point displacements to obtain the total angular point displacement of the production image.
And 105, when the total angular point displacement reaches a target threshold value, determining that the molten material stage is in a collapsed state.
Optionally, in actual implementation, the step includes:
firstly, acquiring angular point displacement acquired in a melting stage;
and secondly, calculating the total displacement of all the corner points on the production image according to the obtained displacement of each corner point.
Thirdly, determining the target threshold value according to the total angular point displacement calculated and obtained in the previous heat.
(1) Sequencing the obtained total displacement of each corner point according to an ascending order, and obtaining a first displacement in a first sequence and a second displacement in a second sequence;
in one possible embodiment, the first ordering is one-fourth and the second ordering is three-quarters, and accordingly, the first shift is determined to be Q1 and the second shift is determined to be Q3.
(2) Calculating the displacement distance between the first displacement and the second displacement;
the displacement distance is IQR 3-Q1.
(3) And determining to obtain the target threshold according to the second displacement and the displacement distance.
The target threshold is: U-Q3 + a IQR. Where a is a coefficient, and in one possible embodiment, a is 1.5.
The quartile is used for calculating the target threshold in real time, and the quartile has certain resistance, so that the noise interference in the material melting process is reduced.
Optionally, when it is determined that the material is in a collapsed state, a prompt may be sent in a preset prompt mode, where the preset prompt mode includes: at least one of voice, indicator light and vibration, and the specific prompting manner is not limited in this embodiment. And after the control system receives the material collapse indication, the crucible position can be automatically and timely lifted, the original temperature gradient is broken, the phenomenon of splashing of silicon solution caused by overlarge local temperature gradient is prevented, the risk of silicon leakage caused by damage of the sharp part of the silicon material to the crucible can be effectively avoided, the automation degree of the material melting process is improved, and the labor cost is reduced.
In conclusion, the production image of the melting stage is obtained in real time; detecting corner information of the molten material according to the production image; determining angular point displacement according to the production image and the angular point information; calculating the total angular point displacement on the production image according to the angular point displacement; and when the total angular point displacement reaches a target threshold value, determining that the molten material stage is in a collapsed state. The problem that detection cannot be performed possibly when judgment is performed according to the number of the angular points in the prior art is solved, and the effect of detecting the collapse state in real time is achieved.
Meanwhile, the angular points are detected through a Shi-Tomasi angular point detection algorithm, so that the angular point detection accuracy is improved, and the subsequent collapse detection accuracy is further improved.
And calculating a target threshold according to each total displacement obtained by calculation, so that whether abnormal displacement occurs can be judged according to the size relation between the total displacement and the target threshold, and the accuracy of the collapse detection is further ensured.
The application also provides a device for detecting the collapsed state, which comprises a memory and a processor, wherein at least one program instruction is stored in the memory, and the processor is used for realizing the method by loading and executing the at least one program instruction.
The present application also provides a computer storage medium having stored therein at least one program instruction, which is loaded and executed by a processor to implement the method as described above.
The technical features of the embodiments described above may be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the embodiments described above are not described, but should be considered as being within the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present invention, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A method for detecting a slump state, the method comprising:
acquiring a production image of a melting stage in real time;
detecting corner information of the molten material according to the production image;
determining angular point displacement according to the production image and the angular point information;
calculating the total angular point displacement on the production image according to the angular point displacement;
and when the total angular point displacement is larger than a target threshold value, determining that the molten material stage is in a collapsed state.
2. The method of claim 1, further comprising:
acquiring angular point displacements acquired in a melting stage;
calculating the total displacement of all the angular points on the production image according to the displacement of each angular point;
and determining the target threshold according to the total angular point displacement obtained by calculation of the current heat.
3. The method of claim 2, wherein determining the target threshold value according to the total corner point displacement calculated by the current heat comprises:
sequencing the total angular point displacement of each acquired frame of picture according to an ascending order, and acquiring a first displacement in a first sequence and a second displacement in a second sequence;
calculating a displacement distance of the first displacement and the second displacement;
and determining to obtain the target threshold according to the second displacement and the displacement distance.
4. The method of claim 1, wherein detecting corner information of the frit from the production image comprises:
corner information in the production image is detected by a Shi-Tomasi corner detection algorithm.
5. The method of claim 1, wherein determining corner displacements from the production image and the corner information comprises:
and determining to obtain the angular point displacement according to the production image and the angular point information by using a Lucas-Kanad algorithm.
6. The method according to claim 5, wherein said determining the corner displacement from the production image and the corner information by Lucas-Kanad algorithm comprises:
reducing the production image for n times to generate a first pyramid image of the production image, wherein the bottom layer of the first pyramid image is the production image; n is a positive integer;
and determining the angular point displacement according to the first pyramid image, the second pyramid image corresponding to the previous frame of production image and the angular point information by using a Lucas-Kanad algorithm.
7. The method of claim 6, wherein determining the corner displacement from the first pyramid image, a second pyramid image corresponding to a previous frame of production image, and the corner information by using a Lucas-Kanad algorithm comprises:
initializing an existing displacement of an n +1 th layer in the second pyramid image;
for the L-th layer of the first pyramid image, obtaining first corner displacement of the L + 1-th layer in the second pyramid image calculated by the Lucas-Kanad algorithm, and calculating the existing displacement of the L-th layer in the second pyramid image according to the first corner displacement and the existing displacement of the L + 1-th layer in the second pyramid image; calculating a second angular point displacement of the L-th layer of the first pyramid image through the Lucas-Kanad algorithm, acquiring the existing displacement of the L-th layer of the first pyramid image according to the existing displacement of the L-th layer of the second pyramid image and the second angular point displacement, and if L is larger than 1, continuing to execute the step; the initial value of L is n;
and if L is 1, determining the determined existing displacement of the 1 st layer in the first pyramid image as the corner displacement.
8. The method of any of claims 1 to 7, wherein said acquiring in real time a production image of a melt phase comprises:
collecting a collected image of the molten material stage through an image sensor;
and preprocessing and gray level conversion are carried out on the collected image, and the preprocessed gray level image is determined as the production image.
9. A slump condition detection apparatus comprising a memory having stored therein at least one program instruction, and a processor for carrying out the method of any one of claims 1 to 8 by loading and executing the at least one program instruction.
10. A computer storage medium having stored therein at least one program instruction which is loaded and executed by a processor to implement the method of any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111159140.XA CN113781474A (en) | 2021-09-30 | 2021-09-30 | Collapse state detection method and device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111159140.XA CN113781474A (en) | 2021-09-30 | 2021-09-30 | Collapse state detection method and device and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113781474A true CN113781474A (en) | 2021-12-10 |
Family
ID=78854772
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111159140.XA Pending CN113781474A (en) | 2021-09-30 | 2021-09-30 | Collapse state detection method and device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113781474A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024066413A1 (en) * | 2022-09-30 | 2024-04-04 | 隆基绿能科技股份有限公司 | Feeding occasion detection method and apparatus, and electronic device and storage medium |
-
2021
- 2021-09-30 CN CN202111159140.XA patent/CN113781474A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024066413A1 (en) * | 2022-09-30 | 2024-04-04 | 隆基绿能科技股份有限公司 | Feeding occasion detection method and apparatus, and electronic device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1202290C (en) | Method and system for controlling growth of a silicon crystal | |
| CN109345593B (en) | Camera posture detection method and device | |
| CN113781474A (en) | Collapse state detection method and device and storage medium | |
| CN104063074B (en) | Information Display Device And Method, | |
| CN108681992A (en) | The image interpolation algorithm of laser facula is measured for detector array method | |
| CN109242019B (en) | Rapid detection and tracking method for optical small target on water surface | |
| CN116071362B (en) | Crystal pulling broken bud detection method, device, computer equipment and storage medium | |
| CN111998910A (en) | Visual measurement method and system for water level of multi-stage water gauge | |
| US20170292204A1 (en) | Manufacturing method and manufacturing system for silicon single crystal | |
| CN112488980A (en) | Molten material state detection method, device and equipment | |
| JP2014501220A (en) | Crystal growth characteristics measurement method using multiple cameras | |
| US20150197874A1 (en) | Device for growing monocrystalline silicon and method for manufacturing the same | |
| CN114369868B (en) | Crystal bar broken line monitoring method, storage medium, terminal and crystal pulling equipment | |
| JPH08133887A (en) | Apparatus for detecting diameter of semiconductor single crystal | |
| CN111932637B (en) | Vehicle body camera external parameter self-adaptive calibration method and device | |
| CN114387248B (en) | Silicon material melting degree monitoring method, storage medium, terminal and crystal pulling equipment | |
| JP2017039628A (en) | Manufacturing method of single crystal and manufacturing apparatus of single crystal | |
| Strauch et al. | Analysis of grain structure evolution based on optical measurements of mc Si wafers | |
| Wang et al. | A novel method for diameter measurement of silicon single crystal | |
| Teng et al. | Onset of sidewise instability and cell–dendrite transition in directional solidification | |
| CN111932624B (en) | Inclination image segmentation-based repose angle detection method | |
| JP2010117259A (en) | Method, device and program for detecting deformed portion on building wall surface | |
| Han et al. | Improved multidomain convolutional neural network for intelligent tracking of Marangoni effect process sequence images | |
| CN114252018A (en) | Crystal diameter detection method, system and computer program product | |
| Wang et al. | A novel measurement system for silicon melt level in Czochralski crystal pullers based on laser triangulation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |