KR102634135B1 - Analysis apparatus and method for cloth - Google Patents

Abstract

The present invention relates to a cloth analysis device and method that enables analysis of the type of cloth using optical characteristics, comprising: a light emitting unit that irradiates light to an analysis object; a light receiving unit that receives reflected light reflected from the analysis object or transmitted light passing through the analysis object; and a control unit that determines the type of fabric of the object to be analyzed according to the characteristics of the light-receiving signal generated from the light-receiving unit.

Description

{Analysis apparatus and method for cloth}

The present invention relates to a cloth analysis device and method, and more specifically, to a cloth analysis device and method that allows the type of cloth to be analyzed using optical characteristics.

In general, the fabric that makes up clothing or various fabrics may be made of fiber materials such as cotton, nylon, wool, acetate, linen, rayon, silk, viscose, polyester, leather, and rayon.

Depending on the type of fabric, the washing method, storage or use method, or production or processing method of clothing or various fabrics varies, so information on the type of fabric must be displayed when selling clothing or various textile products. This is one of the pieces of information.

However, in cases where various types of fabrics are washed or handled in batches, or where it is difficult to check each product tag even if there is a product tag, or when the product tag is lost or difficult to recognize, information about the type of fabric is provided. There were problems in that it was difficult to quickly and accurately obtain the information, and as a result, it was difficult to automate laundry operations and other processes such as cloth production, processing, and handling according to the type of cloth.

The present invention is intended to solve various problems including the problems described above. It is possible to analyze the type of cloth using optical characteristics, and to determine the thickness, degree of adhesion, color, texture, etc. of the cloth using an artificial neural network circuit. We aim to provide a cloth analysis device and method that can improve the accuracy and precision of analysis by calculating accurate results even for variables such as wrinkles. However, these tasks are illustrative and do not limit the scope of the present invention.

A cloth analysis device according to the spirit of the present invention for solving the above problems includes a light emitting unit that irradiates light to an object to be analyzed; a light receiving unit that receives reflected light reflected from the analysis object or transmitted light passing through the analysis object; and a control unit that determines the type of fabric of the object to be analyzed according to the characteristics of the light-receiving signal generated from the light-receiving unit.

Additionally, according to the present invention, the light receiving unit includes: a first light receiving element that receives light in the first wavelength band so as to analyze the absorption spectrum of the first wavelength band in the reflected light or the transmitted light; and a second light receiving element that receives light in the second wavelength band so as to analyze an absorption spectrum of the second wavelength band different from the first wavelength band in the reflected light or the transmitted light.

Additionally, according to the present invention, the first light receiving element includes: a band pass filter that passes light in the first wavelength band; and an image sensor that receives light in the first wavelength band.

In addition, according to the present invention, the control unit performs the light emitting OFF mode and the light receiving OFF mode in the 0th step, the light emitting OFF mode and the light receiving ON mode in the first step, and the light emitting ON mode and the light receiving ON mode in the second step. The ON mode can be performed, the light emitting OFF mode and the light receiving ON mode can be performed in the third step, and the light emitting OFF mode and light receiving OFF mode can be performed in the fourth step.

In addition, according to the present invention, the control unit repeats the 0th step to the 4th step n times (n cycles), accumulates the received optical signals, and converts the accumulated optical pattern for each wavelength into the existing cloth spectrum pattern. By comparing with , the type of fabric of the analysis object can be determined.

In addition, according to the present invention, the control unit is composed of an input layer, a hidden layer, and an output layer, and is divided into existing at least types, thicknesses, adhesion degrees, colors, textures, wrinkles, clothing types, and measurement environments. Learning data about cloth made by selecting any one or a combination thereof is input to the input layer, and the weights of the values filtered in the input layer are used to increase the discrimination accuracy for the learning result in the hidden layer. By adjusting the direction, a deep learning artificial neural network can be used to classify the type of fabric with high accuracy in the output layer.

In addition, according to the present invention, the type of cloth output from the output layer is at least one of cotton, nylon, wool, acetate, linen, rayon, silk, viscose, polyester, leather, rayon, or a combination thereof. It can contain one.

In addition, according to the present invention, a light transmitting window is formed on one side so as to be in contact with or directed to the analysis object, a waveguide space through which the reflected light or the transmitted light can be transmitted is formed inside, and a side of the waveguide space is formed. It may further include a body part on which an inclined surface is formed so that the light emitting part can be installed, and a plurality of the light receiving parts are arranged in a matrix on the bottom of the wave guide space.

Additionally, according to the present invention, a separation barrier may be formed between the light receiving unit arranged in a matrix and neighboring light receiving units to prevent crosstalk.

Meanwhile, the cloth analysis method according to the spirit of the present invention to solve the above problem includes the steps of (a) irradiating light to an analysis object; (b) receiving reflected light reflected from the analysis object or transmitted light passing through the analysis object; and (c) determining the type of fabric of the object to be analyzed according to the characteristics of the light-receiving signal generated from the light-receiving unit.

Additionally, according to the present invention, step (c) includes performing the light emitting OFF mode and the light receiving OFF mode in the 0th step (c-0); (c-1) performing light emitting OFF mode and light receiving ON mode in the first step; (c-2) performing light emitting ON mode and light receiving ON mode in the second step; (c-3) performing light emitting OFF mode and light receiving ON mode in the third step; (c-4) performing light emitting OFF mode and light receiving OFF mode in the fourth step; and (c-5) repeating the 0th step to the 4th step n times (n cycles), accumulating the received optical signals, and analyzing the accumulated optical pattern by wavelength by comparing it with the existing cloth spectrum pattern. It may include a step of determining the type of fabric of the object.

In addition, according to the present invention, the step (c-5) consists of an input layer, a hidden layer, and an output layer, and at least existing clothes by type, thickness, adhesion, color, texture, wrinkle, etc. Learning data about cloth made by selecting one type, measurement environment, or combinations thereof is input to the input layer, and the weight of the values filtered in the input layer is applied to the learning result in the hidden layer. A deep learning artificial neural network can be used to classify the type of fabric with high accuracy in the output layer by adjusting it to increase the discrimination accuracy.

According to an embodiment of the present invention as described above, the type of fabric can be analyzed using optical characteristics, and variables such as thickness, degree of adhesion, color, texture, and wrinkles of the fabric can be measured using an artificial neural network circuit. By calculating accurate results, you can achieve the effect of improving the accuracy and precision of analysis. Of course, the scope of the present invention is not limited by this effect.

1 is a perspective view showing a cloth analysis device according to some embodiments of the present invention.
Figure 2 is a plan view showing the cloth analysis device of Figure 1.
Figure 3 is a cross-sectional view showing the cloth analysis device of Figure 1.
Figure 4 is a plan view showing the circuit board of the cloth analysis device of Figure 1.
Figure 5 is a circuit diagram conceptually showing the cloth analysis device of Figure 1.
FIG. 6 is a time chart showing the cloth analysis process of the cloth analysis device of FIG. 1 over time.
FIG. 7 is a cumulative graph showing light-receiving signals received from light-receiving units of the cloth analysis device of FIG. 1.
FIG. 8 is a graph showing several examples of light spectra for each fabric input to the control unit of the fabric analysis device of FIG. 1.
Figure 9 is a photograph showing several examples of cloth analyzed in the cloth analysis device of Figure 1.
FIG. 10 is a conceptual diagram showing an example of an artificial neural network of the control unit of the cloth analysis device of FIG. 1.
Figure 11 is a flowchart showing a cloth analysis method according to some embodiments of the present invention.
Figure 12 is a flowchart showing step (c) of the cloth analysis method of Figure 11 in more detail.

Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings. However, the present invention is not limited to the embodiments disclosed below, but can be implemented in various different forms. The examples below make the disclosure of the present invention complete, and provide those of ordinary skill in the art with the scope of the invention. It is provided to provide complete information. Additionally, for convenience of explanation, the sizes of components may be exaggerated or reduced in the drawings.

FIG. 1 is a perspective view showing a cloth analysis device 100 according to some embodiments of the present invention, FIG. 2 is a plan view showing the cloth analysis device 100 of FIG. 1, and FIG. 3 is a cloth analysis device of FIG. 1 ( 100), FIG. 4 is a plan view showing the circuit board (PCB) of the cloth analysis device 100 of FIG. 1, and FIG. 5 is a circuit diagram conceptually showing the cloth analysis device 100 of FIG. 1.

First, as shown in FIGS. 1 to 5, the cloth analysis device 100 according to some embodiments of the present invention largely includes a body portion 10, a light emitting portion 20, a light receiving portion 30, and It may include a control unit 40.

For example, the body portion 10 may be a structure in which the light emitting portion 20 and the light receiving portion 30 are installed and has sufficient strength and durability to support them.

More specifically, as shown in FIG. 3, the body portion 10 may be in contact with an analysis object 1 such as cloth, clothing, or various textiles on one side, or may be directed toward the analysis object 1. A transparent window 11, such as a glass window or an acrylic window, is formed, and a waveguide space (A) through which reflected light (L2) or transmitted light can be transmitted is formed inside, and the light emitting part is located on a side of the waveguide space (A). An inclined surface (C) is formed so that (20) can be installed, and a plurality of the light receiving units (30) are arranged on the bottom of the wave guide space (A) in the form of a matrix of n rows and m columns. It may be a structure.

However, the body portion 10 is not necessarily limited to the drawings, and may vary depending on whether it is reflective or transmissive, for example, and may vary depending on the shape, type, or number of the light emitting portion 20 or the light receiving portion 30. It can be formed into many different shapes depending on the shape, type, number, etc.

In addition, for example, the body portion 10 can be installed not only in a washing machine, but also in various places where cloth is produced, processed, handled, or analyzed, and can be applied to washing machines, weaving machines, cloth analyzers, etc. .

In addition, for example, the light emitting unit 20 may be any of various light emitting elements, light emitting lamps, light emitting devices, etc. that irradiate the irradiation light L1 to the analysis object 1.

The light emitting unit 20 may be a light emitting device such as an LED, a halogen lamp, or a discharge tube capable of generating light of a wide wavelength so as to irradiate the irradiation light L1 of a wide wavelength band. For example, , it can generate light in the infrared band as well as light in the visible or ultraviolet band.

Also, for example, the light receiving unit 30 may be a plurality of light receiving devices that receive reflected light L2 reflected from the analysis object 1 or transmitted light passing through the analysis object 1.

More specifically, as shown in FIG. 4, the light receiving unit 30 is arranged in a plurality of n rows and m columns on a circuit board (PCB) on which lamp terminals LT are printed, that is, 4 rows and 4 in the drawing. A total of 16 matrixes can be arranged in rows.

The light receiving unit 30 is used to obtain an overall broadband optical spectrum signal by detecting different wavelength bands of the plurality of light receiving units 30, and absorbs light in the first wavelength band from the reflected light L2 or the transmitted light. A first light receiving element 31 that receives light in the first wavelength band so that the spectrum can be analyzed, and an absorption spectrum of a second wavelength band different from the first wavelength band in the reflected light (L2) or the transmitted light is analyzed. It may include a second light receiving element 32 that receives light in the second wavelength band.

That is, for example, the light receiving unit 30 can receive light in 16 different wavelength bands and analyze the absorption spectrum of the reflected light (L2). The number, arrangement, type, etc. of the light receiving units 30 can be determined. It is not necessarily limited to the drawing, and a wide variety of shapes and types of light-receiving sensors may be applied.

More specifically, for example, as shown in FIGS. 3 and 4, the first light receiving element 31 includes a band pass filter (F) that passes light in the first wavelength band and a band pass filter (F) that passes light in the first wavelength band. It may include an image sensor (S) or a photo sensor that receives light.

Here, all 16 image sensors (S) or photo sensors constituting the 16 light receiving units 30 are the same, and the band pass filters (F) are of 16 types to allow light of different wavelength bands to pass through. may all be different.

However, the number, type, or shape of the light receiving units 30 are not necessarily limited thereto, and a wide variety of light receiving units, types, or shapes may be applied.

In addition, for example, a separation barrier T is formed between the light receiving portion 30 arranged in a matrix and the neighboring light receiving portion 30 to prevent crosstalk caused by optical interference between adjacent light receiving portions, and this separating barrier It is also possible for an inclined surface, etc. to be formed in (T).

Meanwhile, as shown in FIGS. 1 to 5, the control unit 40 includes a microprocessor that determines the type of fabric of the analysis object 1 according to the characteristics of the light-receiving signal generated by the light receiving unit 30; MCU (Micro Controller Unit), computing unit, central processing unit, control device, control panel, circuit board, electronic components, computing chip, integrated circuit, memory device or storage device storing programs, server computer, personal computer, terminal, laptop, It may be various control devices such as smartphones, smart pads, tablets, and smart devices.

More specifically, as shown in FIG. 5, the control unit 40 is connected to the lamps that are the light emitting unit 20, and is connected to a signal processing circuit (ROIC, Read-Out Integrated Circuit) or a multiplex. It may include an MCU (Micro Controller Unit) connected to the light receiving units 30 and 16 channels using a multiplexer (MUX), and other UART (Universal asynchronous receiver/transmitter) or A regulator or power source may be connected.

FIG. 6 is a time chart showing the fabric analysis process of the fabric analysis device 100 of FIG. 1 over time.

As shown in FIG. 6, for example, the control unit 40 performs a light emission OFF mode and a light reception OFF mode for, for example, 600 ms in the 0th step, and performs the light emission OFF mode for, for example, 150 ms in the first step. mode, perform light receiving ON mode, perform light emitting ON mode, light receiving ON mode in the second step, for example for 700 ms, and perform light emitting OFF mode and light receiving ON mode in the third step, for example, for 450 ms. , in the fourth step, for example, for 600 ms, a first cycle of performing the light-emitting OFF mode and the light-receiving OFF mode may be performed.

FIG. 7 is a cumulative graph showing light-receiving signals received from the light-receiving units 30 of the cloth analysis device 100 of FIG. 1.

Next, the control unit 40 repeats the 0th step to the 4th step shown in FIG. 6 n times (n cycles), and generates the optical signals received in the 1st to 3rd steps as shown in FIG. 7. It can be accumulated like a star accumulation graph. Therefore, through this accumulation process, the characteristics of the received optical signal can be revealed more prominently.

FIG. 8 is a graph showing several examples of light spectra for each fabric input to the control unit 40 of the fabric analysis device 100 of FIG. 1.

Next, the control unit 40 converts the accumulated light patterns for each wavelength shown in FIG. 7 to existing fabrics, such as leather wet, leather dry, nylon wet, nylon dry, rayon wet, as shown in FIG. 8. Rayon Dry, Linen Wet, Linen Dry, Cotton Wet, Cotton Dry, Viscose Wet, Viscose Dry, Silk Wet, Silk Dry, Acetate Wet, Acetate Dry, Rayon Wet, Rayon Dry, Poly Wet, Poly Dry, Wool Wet, Wool Dry It is possible to determine the type of fabric of the analysis object (1) by comparing it with existing spectral patterns such as the above.

FIG. 9 is a photograph showing several examples of fabric analyzed by the fabric analysis device 100 of FIG. 1.

Therefore, as shown in FIG. 9, according to the cloth analysis device 100 according to some embodiments of the present invention, cloth such as cotton, nylon, wool, acetate, linen, rayon, silk, viscose, polyester, etc. It can be determined by analyzing by type.

FIG. 10 is a conceptual diagram showing an example of an artificial neural network (ANN) of the control unit 40 of the cloth analysis device 100 of FIG. 1.

As shown in FIG. 10, the control unit 40 consists of an input layer 41, a hidden layer 42, and an output layer 43, and operates at least by type, thickness, adhesion, and color. , learning data on fabric made by selecting one or a combination of texture, wrinkle, clothing type, and measurement environment is input to the input layer 41, and the input layer 41 Deep learning is used to adjust the weight (Wi) of the filtered values in the direction of increasing the discrimination accuracy of the learning result in the hidden layer 42 to classify the type of cloth with high accuracy in the output layer 43. An artificial neural network (ANN) can be used.

Therefore, at least cotton, nylon, wool, acetate, linen, rayon, silk, viscose, polyester, leather, and rayon are selected from the output layer 43 using an artificial neural network (ANN) that learns learning data for various fabrics. Discrimination data for the type of fabric containing any one or a combination thereof can be output.

For example, when the artificial neural network (ANN) uses GENANN, a type of C library, the final calculation time is 20 under the conditions of 1) Input: 16, 2) Layer: 1, 3) Hidden: 45, 4) Output: 9. It was confirmed that it could be reduced to less than a second and determined with high accuracy.

Figure 11 is a flowchart showing a cloth analysis method according to some embodiments of the present invention.

As shown in Figures 1 to 11, the cloth analysis method according to some embodiments of the present invention includes the steps of (a) irradiating irradiation light (L1) to an analysis object (1), (b) the analysis Receiving reflected light (L2) reflected from the object (1) or transmitted light that has passed through the object (1), and (c) analyzing the object (1) according to the characteristics of the light-receiving signal generated from the light receiving unit (30). It may include the step of determining the type of fabric.

Figure 12 is a flowchart showing step (c) of the cloth analysis method of Figure 11 in more detail.

More specifically, as shown in FIG. 12, step (c) includes performing the light emission OFF mode and the light reception OFF mode in the 0th step (c-0), and (c-1) the 0th step. (c-2) performing the light emitting OFF mode and the light receiving ON mode in the first step, (c-2) performing the light emitting ON mode and the light receiving ON mode in the second step, and (c-3) performing the light emitting OFF mode in the third step. , performing the light receiving ON mode, (c-4) performing the light emitting OFF mode and the light receiving OFF mode in the fourth step, and (c-5) performing the 0th step to the 4th step n times (n Cycle) may include repeating, accumulating the received optical signals, and comparing the accumulated optical pattern for each wavelength with the existing cloth spectrum pattern to determine the type of cloth of the analysis object.

Here, the step (c-5) consists of an input layer 41, a hidden layer 42, and an output layer 43, and is divided into at least the existing types, thicknesses, adhesion, colors, textures, Learning data on fabric made by selecting one of wrinkles, clothing types, measurement environments, or combinations thereof is input to the input layer 41, and the values filtered in the input layer 41 A deep learning artificial neural network (ANN) is used to adjust the weights (Wi) of the hidden layer 42 in a direction that increases the discrimination accuracy of the learning result, so that the type of cloth can be classified with high accuracy in the output layer 43. ) can be used.

The present invention has been described with reference to an embodiment shown in the drawings, but this is merely illustrative, and those skilled in the art will understand that various modifications and equivalent other embodiments are possible therefrom. Therefore, the true scope of technical protection of the present invention should be determined by the technical spirit of the attached patent claims.

1: Analytical object
L1: irradiation light
L2: reflected light
10: body part
11: transparent window
A: waveguide space
C: slope
20: light emitting unit
30: light receiving unit
31: first light receiving element
F: Band pass filter
S: image sensor
32: second light receiving element
PCB: circuit board
LT: Lamp terminal
40: control unit
41: Input layer
42: Hidden layer
43: output layer
Wi: weight
ANN: Artificial Neural Network
T: Separating bulkhead
100: Cloth analysis device

Claims (12)

A light emitting unit that irradiates irradiation light to the analysis object;
a light receiving unit that receives reflected light reflected from the analysis object or transmitted light passing through the analysis object; and
a control unit that determines the type of fabric of the object to be analyzed according to the characteristics of the light-receiving signal generated from the light-receiving unit; Including,
The control unit,
In step 0, light emitting OFF mode and light receiving OFF mode are performed,
In the first step, light emission OFF mode and light reception ON mode are performed,
In the second step, light emitting ON mode and light receiving ON mode are performed,
In the third step, light emission OFF mode and light reception ON mode are performed,
In the fourth step, light emitting OFF mode and light receiving OFF mode are performed,
While repeating the 0th step to the 4th step n times (n cycles), the received optical signals are accumulated, and the accumulated optical pattern for each wavelength is compared with the existing cloth spectrum pattern to determine the type of cloth of the object to be analyzed. Distinguish,
It consists of an input layer, a hidden layer, and an output layer, and is comprised of at least one of the existing types, thicknesses, adhesion levels, colors, textures, wrinkles, clothing types, and measurement environments, or combinations thereof. Learning data about cloth made by selecting is input to the input layer, and the weight of the filtered values in the input layer is adjusted in the direction of increasing the discrimination accuracy of the learning result in the hidden layer to determine the type of cloth in the output layer. A cloth analysis device that uses a deep learning artificial neural network to classify with high accuracy. According to claim 1,
The light receiving unit,
a first light receiving element that receives light in the first wavelength band so as to analyze an absorption spectrum of the first wavelength band from the reflected light or the transmitted light; and
a second light receiving element that receives light in the second wavelength band so as to analyze an absorption spectrum of the second wavelength band different from the first wavelength band in the reflected light or the transmitted light;
Including, a cloth analysis device. According to claim 2,
The first light receiving element is,
a band pass filter that passes light in the first wavelength band; and
an image sensor that receives light in the first wavelength band;
Including, a cloth analysis device. delete delete delete According to claim 1,
The type of cloth output from the output layer includes at least one of cotton, nylon, wool, acetate, linen, rayon, silk, viscose, polyester, leather, and rayon, or a combination thereof. Cloth analysis Device. A light emitting unit that irradiates irradiation light to the analysis object;
a light receiving unit that receives reflected light reflected from the analysis object or transmitted light passing through the analysis object;
a control unit that determines the type of fabric of the object to be analyzed according to the characteristics of the light-receiving signal generated from the light-receiving unit; and
A light-transmitting window is formed on one side so as to be in contact with or directed toward the object of analysis, and a waveguide space through which the reflected or transmitted light can be transmitted is formed inside, and the light emitting unit may be installed on a side of the waveguide space. a body portion having an inclined surface formed so as to have a plurality of light receiving portions arranged in a matrix on the bottom of the wave guide space; Including,
A cloth analysis device in which a separation barrier is formed between the light receiving unit arranged in a matrix and neighboring light receiving units to prevent crosstalk. delete (a) irradiating light to the analysis object;
(b) receiving reflected light reflected from the analysis object or transmitted light passing through the analysis object; and
(c) determining the type of fabric of the object to be analyzed according to the characteristics of the light-receiving signal generated from the light-receiving unit; Including,
In step (c),
(c-0) performing light emitting OFF mode and light receiving OFF mode in step 0;
(c-1) performing light emitting OFF mode and light receiving ON mode in the first step;
(c-2) performing light emitting ON mode and light receiving ON mode in the second step;
(c-3) performing light emitting OFF mode and light receiving ON mode in the third step;
(c-4) performing light emitting OFF mode and light receiving OFF mode in the fourth step; and
(c-5) While repeating the 0th step to the 4th step n times (n cycles), the received optical signals are accumulated, and the accumulated optical pattern for each wavelength is compared with the existing cloth spectrum pattern to determine the analysis object. Determining the type of fabric; Including,
In step (c-5),
It consists of an input layer, a hidden layer, and an output layer, and is comprised of at least one of the existing types, thicknesses, adhesion levels, colors, textures, wrinkles, clothing types, and measurement environments, or combinations thereof. Learning data about cloth made by selecting is input to the input layer, and the weight of the filtered values in the input layer is adjusted in the direction of increasing the discrimination accuracy of the learning result in the hidden layer to determine the type of cloth in the output layer. A cloth analysis method that uses a deep learning artificial neural network to classify with high accuracy. delete delete

Images