KR102408754B1 - Method For Virtual Data Generation of Crack Patterns - Google Patents

Abstract

The present invention provides a virtual defect data generation method for virtually generating defect image data of a ceramic surface generated in a ceramic press process, comprising: a first step of inputting a required size and number of two-dimensional models; a second step of generating a defective two-dimensional model according to the data input from the first step; a third step of converting the two-dimensional model generated in the second step into digital data; a fourth step of extracting and classifying defect information from the digital dataized two-dimensional model converted from the third step; and a fifth step of generating the data of the defect information extracted and classified from the fourth step as a data set for each two-dimensional model; It is characterized in that it includes.
Accordingly, a sufficient amount of data may be secured by virtually generating a pattern for a surface defect occurring in the ceramic press process.
In addition, by generating data on surface defects using a random step algorithm, it is possible to secure a defect image in the form of a surface defect pattern that may occur during product manufacturing as much as possible.
In addition, by learning through machine learning based on virtual defect data obtained through big data, information on defect patterns for various properties can be secured, thereby minimizing the defect rate for surface defects occurring in the press process.

Description

Method For Virtual Data Generation of Crack Patterns

The present invention relates to a method for generating virtual defect data.

In the current era of big data analysis and artificial intelligence technology (especially deep learning-related research among machine learning), securing data is an essential condition.

In particular, in the field of image information, for example, companies that have secured big data (eg, Google, Facebook, etc.) are researching and using the latest artificial intelligence technologies such as deep learning.

Image big data is used as training data for artificial intelligence technologies to learn cognition and reasoning capabilities of computer systems. Currently, these artificial intelligence technologies are showing performance that approaches or exceeds human capabilities in areas such as speech recognition, image recognition, and text interpretation. Due to its high cognitive performance, the technology is currently recognized as having an innovative impact that will replace many service jobs.

However, there is a problem in that it is difficult to apply the latest technology because it is difficult to secure a sufficient amount of learning data in industries other than companies or fields where big data can be secured through social networks, especially manufacturing related industries.

On the other hand, lowering the defect rate in the manufacturing process in the ceramic material parts industry is one of the most important issues for reducing cost and period.

Specifically, in order to minimize surface defects (cracks) occurring in the ceramic press process, a study using machine learning is actively conducted to make a defect generation pattern of a result of the press process into big data. In this process, the most difficult and important task is to acquire a large amount of data (eg, defect images), but there was a problem that the actual data collection in the field had to be very limited.

Republic of Korea Patent Registration No. 10-1964282 (published on April 1, 2019)

An object of the present invention is to provide a method for generating virtual defect data, which was created to solve the above problem, and in which a sufficient amount of data can be secured by virtually generating a pattern for a surface defect occurring in a ceramic press process. is to do

Another object of the present invention is to provide a method for generating virtual defect data that can secure a defect image in a form as close as possible to a surface defect pattern that may occur during product manufacturing by generating data on surface defects using a random step algorithm. is to do

Another object of the present invention is to minimize the defect rate for surface defects occurring in the press process by securing information on defect patterns for various properties by learning through machine learning based on virtual defect data obtained through big data. It is to provide a method for generating virtual defect data that can

Other detailed objects of the present invention will be clearly grasped and understood by experts or researchers in the technical field through the detailed contents described below.

According to the present invention, there is provided a virtual defect data generation method for virtually generating defect image data of a ceramic surface generated in a ceramic press process, comprising: a first step of inputting a required size and number of two-dimensional models; a second step of generating a defective two-dimensional model according to the data input from the first step; a third step of converting the two-dimensional model generated in the second step into digital data; a fourth step of extracting and classifying defect information from the digital dataized two-dimensional model converted from the third step; and a fifth step of generating the data of the defect information extracted and classified from the fourth step as a data set for each two-dimensional model; It can be achieved by a virtual defect data generation method comprising a.

Here, the first step may include: Step 1-1 of determining the size of the two-dimensional model; Step 1-2 of determining the number of defects per model and the length of each defect; Steps 1-3 of setting the maximum direction change angle (θmax) for each defect; and steps 1-4 determining the number of required models; includes

In addition, the second step may be provided as a step of generating the data input from the first step into a defective two-dimensional model using a random walk algorithm.

Here, the second step includes: a 2-1 step of selecting arbitrary coordinates to generate a first defect with respect to a given model; a step 2-2 of generating coordinates moved by an arbitrary length (ℓ) in an arbitrary direction (θ) from the arbitrary coordinates based on the one model size; a step 2-3 of repeating the step 2-2 within the maximum direction change angle based on the one model size to obtain a coordinate set representing the first defect; Steps 2-4 of selecting arbitrary coordinates to generate an n-th defect with respect to a given model; a step 2-5 of obtaining a coordinate set representing the n-th defect through repetition of steps 2-2 to 2-3 with the coordinates selected in step 2-4; and a step 2-6 of obtaining a coordinate set by repeating steps 2-4 to 2-5 until the number of defects determined in step 1-2 is satisfied; includes

In addition, each coordinate of the coordinate set expressing the n-th defect may be set such that a distance from the coordinate set expressing the n-1 th defect is greater than the arbitrary length ℓ.

Here, the third step may include: a 3-1 step of converting the real number type coordinate set obtained through the second step into the closest integer type coordinate set; and a 3-2 step of obtaining a bitmap mapped to 0 or 1 by displaying the coordinate set converted in step 3-1 as 1 on a bitmap corresponding to the model size and having total coordinates set to 0; includes

Meanwhile, according to the present invention, the above object can also be achieved by a computer-readable recording medium in which computer program codes for performing the method for generating virtual defect data are stored.

According to the present invention, a sufficient amount of data can be secured by virtually generating a pattern for a surface defect occurring in a ceramic press process.

In addition, by generating data on surface defects using a random step algorithm, it is possible to secure a defect image in the form of a surface defect pattern that may occur during product manufacturing as much as possible.

In addition, by learning through machine learning based on virtual defect data obtained through big data, information on defect patterns for various properties can be secured, thereby minimizing the defect rate for surface defects occurring in the press process.

1 is a flowchart of a method for generating virtual defect data according to the present invention;
2 (a), (b) are photographs showing various defect (crack) patterns that occur on the surface of a ceramic product in the ceramic press process,
3 is a representation of the basic principle of a random step algorithm in a method for generating virtual defect data according to an embodiment of the present invention;
4 shows a defect image generated by a method for generating virtual defect data according to an embodiment of the present invention;
5 is a table in which defect-related information secured by a method for generating virtual defect data according to an embodiment of the present invention is classified.

Since the present invention can have various changes and can have various forms, specific embodiments are illustrated in the drawings and described in detail in the text. However, this is not intended to limit the present invention to the specific disclosed form, it should be understood to include all modifications, equivalents and substitutes included in the spirit and scope of the present invention.

The terms used in the present application are only used to describe specific embodiments, and are not intended to limit the present invention. The singular expression includes the plural expression unless the context clearly dictates otherwise. In the present application, terms such as "comprises" or "comprising" are intended to designate that a feature, number, step, operation, component, part, or combination thereof described in the specification is present, but one or more other features It should be understood that it does not preclude the possibility of the presence or addition of numbers, steps, operations, elements, parts, or combinations thereof.

In addition, the defect data generation method described in the present application may be implemented as a computing device having a control unit, a calculation unit, etc., but is not limited thereto, and the subject performing each specific step is each step for rapid simulation and calculation. It may be a separate configuration that performs the

Meanwhile, unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by those of ordinary skill in the art to which the present invention belongs.

Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and should not be interpreted in an ideal or excessively formal meaning unless explicitly defined in the present application. does not

With reference to the drawings below, preferred embodiments of the present invention will be described in more detail.

1 is a flowchart of a method for generating virtual defect data according to the present invention, and FIGS. 2 (a) and (b) are photographs showing various defect (crack) patterns occurring on the surface of a ceramic product in a ceramic press process, and FIG. 3 represents the basic principle of the random step algorithm in the method for generating virtual defect data according to an embodiment of the present invention, and FIG. 4 shows a defect image generated by the method for generating virtual defect data according to an embodiment of the present invention. , FIG. 5 is a table in which defect-related information secured by a method for generating virtual defect data according to an embodiment of the present invention is classified.

1 to 5 , the virtual defect data generating method according to the present invention is a virtual defect data generating method for virtually generating defect image data of a ceramic surface generated in a ceramic press process, and the size of a required two-dimensional model and a first step (S10) of inputting the number of models; a second step (S20) of generating a defective two-dimensional model according to the data input from the first step (S10); a third step (S30) of converting the two-dimensional model generated in the second step (S20) into digital data; a fourth step (S40) of extracting and classifying defect information in the digital dataized two-dimensional model converted from the third step (S30); and a fifth step (S50) of generating the data of the defect information extracted and classified from the fourth step (S40) as a data set for each two-dimensional model; may include.

Here, the first step (S10) includes: a 1-1 step of determining the size of the two-dimensional model; Step 1-2 of determining the number of defects per model and the length of each defect; Steps 1-3 of setting the maximum direction change angle (θmax) for each defect; and steps 1-4 determining the number of required models; includes

In addition, the second step (S20) may be prepared as a step of generating the data input from the first step (S10) into a defective two-dimensional model using a random walk algorithm.

Here, the second step (S20) includes: a 2-1 step of selecting arbitrary coordinates to generate a first defect with respect to a given model; a step 2-2 of generating coordinates moved by an arbitrary length (ℓ) in an arbitrary direction (θ) from the arbitrary coordinates based on the one model size; a step 2-3 of repeating the step 2-2 within the maximum direction change angle based on the one model size to obtain a coordinate set representing the first defect; Steps 2-4 of selecting arbitrary coordinates to generate an n-th defect with respect to a given model; a step 2-5 of obtaining a coordinate set representing the n-th defect through repetition of steps 2-2 to 2-3 with the coordinates selected in step 2-4; and a step 2-6 of obtaining a coordinate set by repeating steps 2-4 to 2-5 until the number of defects determined in step 1-2 is satisfied; includes

In addition, each coordinate of the coordinate set expressing the n-th defect may be set such that a distance from the coordinate set expressing the n-1 th defect is greater than the arbitrary length ℓ.

Here, the third step (S30) includes: a 3-1 step of converting the real number coordinate set obtained through the second step into the closest integer coordinate set; and a 3-2 step of obtaining a bitmap mapped to 0 or 1 by displaying the coordinate set converted in step 3-1 as 1 on a bitmap corresponding to the model size and having total coordinates set to 0; includes

Meanwhile, the present invention may be implemented by a computer-readable recording medium storing computer program codes for performing the above-described method for generating virtual defect data.

Hereinafter, the steps and specific processes of the method for generating virtual defect data according to the present invention will be described by way of example.

Step 1 (S10) - Input setting for the size and required number of the two-dimensional model

1. Determination of size of two-dimensional model - Lx, Ly (length in x and y direction)

2. Determine the number of defects per model (Nc) and the length of each defect (Lc)

Ex) Nc = 1~5, Lc<Lx, Lc<Ly

Lx=640, Ly=480, Lc=50, 100, 150, 200

3. Determination of the maximum turning angle θmax

4. Determine the total number of required models and prepare to run model generation

Example) If Nr (example 5) models are created for each number of defects for each length,

Total number of models (Ntotal)=# of Nc * # of Lc * Nr = 5 * 4 * 5 =100

where # of Nc is the number of defects (Nc) and # of Lc is the number of lengths.

Step 2 (S20) - Create a defective two-dimensional model

1. To generate the first defect for a given model (given the number of defects Nc and the length Lc of each defect), select a random point (P1x_0, P1y_0)

2. 0<P1x_0<Lx, 0<P1y_0<Ly

3. Move by any length (ℓ) in any direction (θ) - Coordinates (P1x_1, P1y_1)

4. Repeat the process of 3. as much as Lc/ℓ (eg ℓ = 1) (the limit that satisfies the conditions of 2., within the maximum direction change angle - θ < θmax)

5. Coordinates representing one defect (P1x_0, P1y_0), (P1x_1, P1y_1)… (P1x_q, P1y_q) acquisition (where q<=Lc)

6. Pick a random point (P2x_0, P2y_0) to create the second fault.

Here, proceed with steps 2 to 4.

In each process, the distance from the coordinates (P2x_i, P2y_i) to all points of the previous defect ((P1x_0, P1y_0), (P1x_1, P1y_1)…(P1x_q, P1y_q)) must be greater than ℓ, if the distance is less than ℓ end

7. Coordinates representing the second defect (P2x_0, P2y_0), (P2x_1, P2y_1)… Acquire (P2x_q, P2y_q) (where q<=Lc)

8. Repeat steps 6 to 7 as many as the number of defects (Nc).

Step 3 (S30) - Convert each model into digital data

1. The data obtained from the process S20 above for one model are real number coordinates

Coordinates for n defects:

(P1x_0, P1y_0), (P1x_1, P1y_1)… (P1x_q, P1y_q)

(P2x_0, P2y_0), (P2x_2, P2y_1)… (P2x_q, P2y_q)

…

(Pnx_0, Pny_0), (Pnx_1, Pny_1)… (Pnx_q, Pny_q)

0<x-coordinate<Lx, 0<y-coordinate<Ly

ex) Here, Lx is 600, Ly is 480

2. Select the nearest integer for all coordinates

0<=Px<600, 0<=Py<480

Px=P1x_0,… P1x_q,… Pnx_0,… Pnx_q, Py=P1y_0,… P1y_q,… Pny_0,… Pny_q

3. Convert to bitmap

Prepare a bitmap of 600X480 consisting of all 0s, transform each coordinate (Px, Py) obtained in 2. above into 1, thus obtaining a bitmap of 600X480 consisting of 0 or 1

Step 4 (S40) - Classification of defect information such as the number of defects, the length of the defects, and the location coordinates

1. Each model is classified by the number of defects planned in S10, etc.

2. The length of each defect is the number of steps of each defect in S30.

3. The total length is the sum of the lengths of all defects.

4. The location of each defect means the coordinates of the rectangle including each defect (eg upper left and lower right coordinates).

5. Classify data according to the various criteria listed above (see Fig. 5)

Step 5 (S50) - Create big data data set for the required number

1. Classify data by various properties set in S10 (refer to Fig. 5) and generate big data for it

As described above, in the method for generating virtual defect data according to the present invention, a sufficient amount of data can be secured by virtually generating a pattern for a surface defect occurring in a ceramic press process.

In addition, by generating data on surface defects using a random step algorithm, it is possible to secure a defect image in the form of a surface defect pattern that may occur during product manufacturing as much as possible.

In addition, by learning through machine learning based on virtual defect data obtained through big data, information on defect patterns for various properties can be secured, thereby minimizing the defect rate for surface defects occurring in the press process.

In the foregoing detailed description of the present invention, although it has been described with reference to preferred embodiments of the present invention, those skilled in the art or those having ordinary knowledge in the art will have the spirit of the present invention described in the claims to be described later. And it will be understood that various modifications and variations of the present invention can be made without departing from the technical scope.

Claims (7)

A virtual defect data generation method for virtually generating defect image data of a ceramic surface generated in a ceramic press process, the method comprising:
a first step of inputting the required size of the two-dimensional model and the number of models;
a second step of generating a defective two-dimensional model according to the data input from the first step;
a third step of converting the two-dimensional model generated in the second step into digital data;
a fourth step of extracting and classifying defect information from the digital dataized two-dimensional model converted from the third step; and
A fifth step of generating the data of the defect information extracted and classified from the fourth step as a data set for each two-dimensional model,
The first step is
Step 1-1 determining the size of the two-dimensional model;
Step 1-2 of determining the number of defects per model and the length of each defect;
Steps 1-3 of setting the maximum direction change angle (θmax) for each defect; and
Steps 1-4 of determining the number of models required;
The second step is a step of generating the data input from the first step into a defective two-dimensional model using a random walk algorithm,
The second step is
Step 2-1 of selecting arbitrary coordinates to generate a first defect with respect to a given model;
a step 2-2 of generating coordinates moved by an arbitrary length (ℓ) in an arbitrary direction (θ) from the arbitrary coordinates based on the one model size;
a step 2-3 of obtaining a coordinate set expressing the first defect by repeating steps 2-2 within the maximum direction change angle based on the one model size;
Steps 2-4 of selecting arbitrary coordinates to generate an n-th defect with respect to a given model;
a step 2-5 of obtaining a coordinate set representing the n-th defect through repetition of steps 2-2 to 2-3 with the coordinates selected in step 2-4; and
Virtual defect data comprising steps 2-6 to obtain a coordinate set by repeating steps 2-4 to 2-5 until the number of defects determined in step 1-2 is satisfied How to create. delete delete delete According to claim 1,
Each coordinate of the coordinate set representing the nth defect has a distance from the coordinate set representing the n−1th defect is greater than the arbitrary length (ℓ). 6. The method of claim 5,
The third step is
a 3-1 step of converting the real number coordinate set obtained through the second step into the closest integer coordinate set; and
a 3-2 step of obtaining a bitmap mapped to 0 or 1 by displaying the coordinate set converted in step 3-1 as 1 on a bitmap corresponding to the model size and having total coordinates set to 0; characterized in that it comprises
How to generate virtual defect data. A computer-readable recording medium storing computer program codes for performing the method for generating virtual defect data according to any one of claims 1 and 5 to 6 .

Images