KR102427597B1 - Fine dust detecting solution and system by computing saturation residual based on AI - Google Patents

Abstract

The present invention relates to a fine dust reading solution and system through artificial intelligence-based chroma residual calculation. acquisition stage; a continuous image extraction step of extracting a plurality of continuous still images from the moving picture; a chroma data calculation step of calculating a chroma data set for each of the still images; a residual calculation step of calculating residual data between two consecutive chroma data sets among the plurality of chroma data sets; and a fine dust level calculation step of calculating the fine dust level based on the residual data.
Accordingly, it is possible to economically measure the level of fine dust in various places.

Description

Fine dust detecting solution and system by computing saturation residual based on AI}

The present invention relates to a solution and system for reading fine dust through artificial intelligence-based chroma residual calculation, and more particularly, through artificial intelligence-based chroma residual calculation that can economically measure fine dust level using an image It relates to a fine dust reading solution and system.

Recently, fine dust has become a social issue, and various media such as news also provide the concentration of fine dust for each region.

Existing fine dust measurement methods include indirect measurement using the physical properties of radiation or light (beta absorption method or light scattering method, etc.) and measuring the mass of fine dust with a scale (weight concentration method, etc.). , expensive measuring equipment is required to measure fine dust in this way.

On the other hand, fine dust measurement is measured for each administrative district unit, but even in the same administrative district, the actual measurement value of fine dust may differ depending on the characteristics of the terrain or the density of buildings.

Therefore, there is a need for a method that can measure the concentration of fine dust in various areas without expensive fine dust measuring equipment.

US 10-2020-009707 A

Accordingly, it is an object of the present invention to solve such conventional problems, and a solution and system for reading fine dust through artificial intelligence-based chroma residual calculation that can economically measure fine dust levels at various locations by using images is to provide.

The problem to be solved by the present invention is not limited to the above-mentioned problems, and other problems not mentioned will be clearly understood by those skilled in the art from the following description.

The above object, according to the present invention, a video acquisition step of acquiring a video indoors or outdoors; a continuous image extraction step of extracting a plurality of continuous still images from the moving picture; a chroma data calculation step of calculating a chroma data set for each of the still images; a residual calculation step of calculating residual data between two consecutive chroma data sets among the plurality of chroma data sets; and a fine dust degree calculation step of calculating the fine dust level based on the residual data.

In the step of calculating the degree of fine dust, the degree of fine dust may be calculated by being divided into classes.

In the step of calculating the degree of fine dust, the degree of fine dust may be calculated through a pair of influence variables comprising two variables in which the dominance of the frequency of each other in different classes is reversed.

The influencing variable pair may be determined as a value when the result of Equation 1 below becomes the maximum.

[Equation 1]

: 두 변수가 x, y일 때 미세먼지 정도가 낮은 클래스에서의 관심도와 미세먼지 정도가 높은 클래스에서의 관심도 차이의 절대값,

: 두 변수가 x, y일 때 미세먼지 정도가 낮은 클래스에서의 관심도,

: 두 변수가 x, y일 때 미세먼지 정도가 높은 클래스에서의 관심도)(

: The absolute value of the difference in interest in the class with a low level of fine dust and the level of interest in the class with a high level of fine dust when the two variables are x and y,

: When the two variables are x and y, the degree of interest in the class with a low level of fine dust,

: Interest in a class with a high degree of fine dust when two variables are x and y)

When the result value of Equation 1 is the same, two variables in the case where the sum of the frequencies of the two variables in each class is the maximum may be determined as the influencing variable pair.

The pair of influence variables may be derived by learning the residual data through an artificial intelligence analysis model.

In the step of calculating the degree of fine dust, the degree of fine dust may be calculated by using an odd number of pairs of three or more influencing variables.

In the step of calculating the degree of fine dust, the degree of fine dust calculated through the pair of influence variables is determined again through a second pair of influencing variables comprising two second variables in which the dominance of the frequency between each other in different subclasses is reversed. It may include a detailed class calculation step.

The fine dust reading solution through the artificial intelligence-based chroma residual calculation according to the present invention is performed between the residual calculation step and the fine dust degree calculation step, and the residual data filtering that excludes a component greater than or equal to a predetermined value among the components of the residual data It may include further steps.

According to another embodiment of the present invention, a video acquisition unit for acquiring a video indoors or outdoors; a continuous image extraction unit for extracting a plurality of continuous still images from the moving picture; a chroma data calculator for calculating a chroma data set of each of the still images; a residual calculator configured to calculate residual data between two consecutive chroma data sets among the plurality of chroma data sets; and a fine dust degree calculation unit for calculating the degree of fine dust based on the residual data.

The fine dust degree calculation unit calculates the fine dust level by dividing it into classes, and the fine dust level may be calculated using a pair of influence variables comprising two variables in which the dominance of the frequency of each other in different classes is reversed.

According to the fine dust reading solution through artificial intelligence-based chroma residual calculation according to the present invention, the level of fine dust can be economically measured. make it known

In addition, it is possible to reduce the measurement time by converting each still image into chroma data during the fine dust level measurement process.

In addition, in the process of measuring the level of fine dust, the degree of fine dust can be easily determined by using the pair of influence variables.

1 is a flowchart of a fine dust reading solution through artificial intelligence-based chroma residual calculation according to the present invention;
2 is an explanatory diagram related to the chroma data calculation step constituting the fine dust reading solution through artificial intelligence-based chroma residual calculation according to the present invention;
3 is an explanatory diagram of an analysis model implementing an influence variable pair algorithm used in a fine dust reading solution through artificial intelligence-based chroma residual calculation according to the present invention;
4 is an explanatory diagram related to the detailed class calculation step constituting the fine dust reading solution through artificial intelligence-based chroma residual calculation according to the present invention;
5 is a schematic configuration diagram of a fine dust reading system through artificial intelligence-based chroma residual calculation according to the present invention.

Hereinafter, specific embodiments of the present invention will be described in detail with reference to the drawings.

1 is a flowchart of a fine dust reading solution through artificial intelligence-based chroma residual calculation according to the present invention.

The fine dust reading solution through artificial intelligence-based chroma residual calculation according to the present invention includes a video acquisition step (S10), a continuous image extraction step (S20), a saturation data calculation step (S30), a residual calculation step (S40) and fine dust It includes a degree calculation step (S50).

In the video acquisition step (S10), a video of a place where the level of fine dust is to be measured is acquired.

A moving picture may have a length of about 5 seconds and may be captured by a digital video camera or a mobile phone. In the present invention, the fine dust reading solution through the calculation of the chroma residual based on artificial intelligence can be applied both indoors and outdoors.

In the continuous image extraction step (S20), a plurality of continuous still images are extracted from the video captured in the video acquisition step (S10).

A moving picture is actually made up of several still images, each appearing in sequence for a short period of time, with minute changes. In the continuous image extraction step (S20), each still image constituting the moving image is extracted. For example, when a moving picture is recorded at 15 fps, 15 frames per second may be extracted from a still image.

Light can be scattered by colliding with fine dust, and the probability of light colliding with fine dust increases as the concentration of fine dust increases. And although it is not easy to confirm with the naked eye, light scattering by fine dust can be captured through a camera.

However, since fine dust is not still in the air, light is scattered differently by the fine dust from moment to moment, and light in a different state is captured in each still image extracted from the video. The higher the concentration of fine dust, the greater the difference between each still image.

In the chroma data calculation step S30, a chroma data set for each still image extracted in the continuous image extraction step S20 is calculated. That is, the saturation data of each pixel constituting the still image is calculated. Since one still image is composed of many pixels, the saturation data set can be calculated for each still image.

As shown in FIG. 2 , the light color of each pixel in a digital image may be expressed as RGB (Red, Green, Blue), and each channel of Red, Green, and Blue may have a value between 0 and 255. have.

When each still image is converted into a chroma data set, since each pixel of the image can be expressed as one channel expressing chroma, the chroma data set has a reduced data size compared to that of a still image. Accordingly, it is possible to shorten the image processing and analysis time.

Since there are many still images extracted from moving images, and each still image consists of many pixels, conversion to a saturation data set can have a very effective effect.

In the saturation data calculation step S30, the saturation data of each pixel of the still image may be calculated by the following method.

인 경우 채도 S는 0이 되고,

≠0 인 경우 채도 S는

가 된다.

, the saturation S becomes 0,

If ≠ 0, the saturation S is

becomes

는

이며,

이고

이다.From here,

Is

is,

ego

to be.

,

,

이다.and

,

,

to be.

In the residual calculation step S40, residual data between two consecutive chroma data sets among a plurality of chroma data sets is calculated.

As described above, different states of light are captured in each of the still images due to light scattering by fine dust, and these states of light are also reflected in each of the chroma data sets calculated from the still images. Therefore, a difference, that is, a residual, occurs between two consecutive chroma data sets.

The residual data may be calculated by the following equation. That is, it can be calculated by finding a difference in values of pixels at the same location in two consecutive chroma data sets (a t-th chroma data set and a t+1 th chroma data set).

: i행, j열의 잔차,

: t번째 채도 데이터 세트에서 i행, j열의 값,

: t+1번째 채도 데이터 세트에서 i행, j열의 값)(

: the residual in row i and column j,

: the value of row i and column j in the tth chroma data set,

: Values in row i and column j in the t+1th saturation data set)

As described above, the higher the fine dust concentration, the greater the difference between each still image.

For example, 75 still images may be extracted from a 5-second video shot at 15 fps, and 74 residual data may be calculated from 75 saturation data sets generated by converting each still image.

In the fine dust level calculation step S50, the fine dust level is calculated based on the residual data.

When the value of the residual is small, it can be determined that the concentration of fine dust is low, and when the value of the residual is large, it can be determined that the concentration of fine dust is high.

The reference data for judging the degree of fine dust can be made by subjecting moving images shot at different fine dust concentrations to still image calculation, chroma data set calculation, and residual data calculation processes, and each of these processes is repeated multiple times. By iteratively performing, the accuracy of the reference data can be improved.

According to the fine dust reading solution through the artificial intelligence-based chroma residual calculation of the present invention, it is economical because it is possible to measure the level of fine dust using a general digital video camera without expensive measuring equipment. Since the level of fine dust can be measured in several places, it is possible to know the local level of fine dust.

Using CCTV as a camera to acquire video, it is also possible to measure the level of fine dust without installing additional hardware equipment.

In the fine dust level calculating step S50, the fine dust level may be calculated by dividing the class into classes.

In this case, the good or bad of fine dust can be intuitively grasped.

A criterion for classifying a class may be determined by applying a binary classification algorithm to a plurality of criterion residual data.

The class of fine dust level can be divided into two classes of, for example, 'good' and 'bad', or into four classes of 'good', 'normal', 'bad', and 'very bad'. .

In the fine dust degree calculation step ( S50 ), the degree of fine dust may be calculated through a pair of influence variables consisting of two variables in which the dominance of the frequency of each other in different classes is reversed.

In this case, it is possible to easily calculate the class of fine dust by judging which of the two variables constituting the pair of influence variables has the highest frequency.

Each residual data can be judged through a pair of influence variables, and a class of the degree of fine dust calculated with a relatively high frequency can be determined as the final result.

As an algorithm for deriving an influence variable pair, an influence variable pair algorithm may be used. By the influencing variable pair algorithm, the influencing variable pair may be determined as a value when the result value of Equation 1 below becomes the maximum.

: 두 변수가 x, y일 때 미세먼지 정도가 낮은 클래스에서의 관심도와 미세먼지 정도가 높은 클래스에서의 관심도 차이의 절대값,

: 두 변수가 x, y일 때 미세먼지 정도가 낮은 클래스에서의 관심도,

: 두 변수가 x, y일 때 미세먼지 정도가 높은 클래스에서의 관심도)(

: The absolute value of the difference in interest in the class with a low level of fine dust and the level of interest in the class with a high level of fine dust when the two variables are x and y,

: When the two variables are x and y, the degree of interest in the class with a low level of fine dust,

: Interest in a class with a high degree of fine dust when two variables are x and y)

In this case, the degree of interest in each class may be calculated by the following equation.

,

,

: 해당 클래스(c)에서 변수

가

보다 낮을 확률,

: 변수 x의 빈도수,

: 변수 y의 빈도수)(

: Variable in the corresponding class (c)

go

less likely,

: the frequency of the variable x,

: frequency of variable y)

For example, if there is reference residual data with the following values for good (c=low) and bad (c=high) fine dust,

The frequencies for the variables constituting the baseline residual data of each class are as follows.

< Frequency distribution table of variables constituting the standard residual data of each class >

For example, based on the frequency distribution table above, if the two variables x and y are +1 and +5, the interest in the class with the low fine dust level is 0, and the interest in the class with the high fine dust level is 1 becomes this

Therefore, when the two variables x and y are +1 and +5, the absolute value of the difference between the interest in a class with a low degree of fine dust and a class with a high degree of fine dust is maximum at 1.

Therefore, two variables +1 and +5 can be set as a pair of influence variables.

Based on this pair of influence variables, when the frequency of variable +1 in the residual data to be judged is high, the degree of fine dust can be judged to be low, and when the frequency of +5 is high, the degree of fine dust is judged to be high can do.

On the other hand, there may be a case in which the absolute value of the difference between the interest in a class with a low degree of fine dust and the degree of interest in a class with a high degree of fine dust is the same.

For example, based on the frequency distribution table above, in addition to the case where the two variables x and y are +1 and +5, even when the two variables x and y are +2, +5, the degree of interest in the class with a low level of fine dust is It becomes 0 and the interest in the class with high fine dust level becomes 1,

The absolute value of the difference between the interest in a class with a low level of fine dust and the level of interest in a class with a high degree of fine dust is 1.

In this case, the accuracy of calculating the degree of fine dust can be improved by setting the two variables in the case where the sum of the frequencies of the two variables in each class is the maximum as a pair of influence variables.

The influence variable pair algorithm can be implemented by an analysis model to which artificial intelligence technology is applied. The analysis model may be various artificial intelligence models such as artificial neural network (ANN), machine learning, deep learning, and the like.

3 shows an example of an analysis model implementing the influence variable pair algorithm. The analysis model may include an input layer (input), a hidden layer (hidden, hidden 2), and an output layer (output). The number of hidden layers may be different from that shown.

In the input layer, values of components constituting residual data when fine dust is good (c=low) and values of components constituting residual data when fine dust is bad (c=high) are respectively input.

The hidden layer (hidden, hidden 2) learns the component values of the residual data of each class, so that the absolute value of the difference between the interest when fine dust is good (c = low) and when the fine dust is bad (c = high) is the maximum. Find the two components (variables) that become

And the two components obtained in this way are output through the output layer to form an influence variable pair.

The pair of influence variables can be derived through the above analysis model path, but it is not limited thereto, and the pair of influence variables can also be derived using other analysis models.

In the fine dust level calculation step ( S50 ), the fine dust level may be calculated using an odd number of pairs of three or more influencing variables.

In this case, the class of the degree of fine dust calculated through the number of pairs of influence variables exceeding the majority can be determined as the result of the final level of fine dust. For example, if the degree of fine dust is calculated using three pairs of influence variables, the degree of fine dust is judged to be high through two pairs of influence variables, and the degree of fine dust is determined through one pair of influence variables. If it is determined that the level of fine dust is low, it can be finally determined that the degree of fine dust is high.

Since the method of calculating the degree of fine dust through an influencing variable pair is based on probability, it is possible to increase the accuracy of the calculation by calculating the degree of fine dust through several influencing variable pairs.

The fine dust level calculation step S50 may include a detailed class calculation step.

In the detailed class calculation step, as shown in FIG. 4 , the degree of fine dust calculated through the pair of influence variables is divided into detailed classes and calculated. Classification into subclasses may be made through a second influencing variable pair consisting of two second variables in which the predominance of frequencies is reversed in different subclasses. That is, the method of classifying the degree of fine dust into detailed classes is different from the method of calculating the degree of fine dust through the pair of influence variables, except that the second variable constituting the second pair of influence variables is different from the variable constituting the pair of influence variables. same.

It is natural that the second pair of influencing variables applicable to the case where the degree of fine dust is judged to be high and the case where the degree of fine dust is judged to be low by the pair of influence variables are different from each other.

If the degree of fine dust was classified into 'low' and 'high' by the influencing variable pair, 'good', 'normal', 'bad', 'very bad' by the second influencing variable pair ' can be further subdivided.

A residual data filtering step may be further performed between the residual calculating step S40 and the fine dust level calculating step S50.

In the residual filtering step, an operation of excluding a component having a value greater than or equal to a predetermined value among residual data components is performed.

In addition to scattering light by fine dust, moving objects can be captured in the video, and the change in light caused by the movement of an object is large enough to be seen with the naked eye. will cause And the residuals due to the movement of these objects act as noise in calculating the degree of fine dust.

Accordingly, it is possible to increase the accuracy of calculating the degree of fine dust by excluding a component having a size greater than or equal to a predetermined value among residual data components.

Hereinafter, the fine dust reading system 1 through artificial intelligence-based chroma residual calculation according to the present invention will be described. 5 is a schematic configuration diagram of a system 1 of the present invention.

While explaining the fine dust reading system 1 through artificial intelligence-based chroma residual calculation according to the present invention, detailed description of the parts mentioned in the description of the fine dust reading solution through artificial intelligence-based chroma residual calculation will be omitted. do.

The fine dust reading system 1 through artificial intelligence-based chroma residual calculation according to the present invention includes a video acquisition unit 10, a continuous image extraction unit 20, a chroma data calculation unit 30, and a residual calculation unit 40. and a fine dust degree calculation unit 50 .

The video acquisition unit 10 acquires a video indoors or outdoors. The moving picture acquisition unit 10 may be, for example, a digital video camera or a mobile phone.

The continuous image extracting unit 20 extracts a plurality of continuous still images from the moving image obtained through the moving image obtaining unit 10 . Light in different states may be captured in each still image due to scattering of light by fine dust.

The chroma data calculator 30 calculates a chroma data set from each of the still images. Compared to a still image expressed in RGB (Red, Green, Blue), the chroma data set has a small size.

The residual calculator 40 calculates residual data between two successive chroma data sets among a plurality of chroma data sets. The degree of light scattering varies according to the degree of fine dust, and the residual between two consecutive chroma data sets varies according to the degree of light scattering. That is, the residual data varies depending on the level of fine dust.

The fine dust level calculating unit 50 calculates the fine dust level based on the residual data. Since the residual data depends on the level of fine dust, it is possible to calculate the level of fine dust from the residual data.

The fine dust degree calculation unit 50 calculates the fine dust level by dividing it into classes, and the fine dust level can be calculated using a pair of influence variables consisting of two variables in which the dominance of the frequency of each other in different classes is reversed. .

In this case, the class of fine dust can be easily calculated by determining which of the two variables constituting the pair of influence variables has the highest frequency.

The fine dust degree calculation unit 50 includes an influence variable pair derivation unit (not shown) that derives an influence variable pair through an analysis model to which artificial intelligence technology is applied, and analyzes the reference residual data from numerous videos. can be derived. And the influencing variable pair derived from the influencing variable pair derivation unit is used to calculate the degree of fine dust.

Configurations of the fine dust reading system 1 through artificial intelligence-based chroma residual calculation may be implemented through one or more hardware, or may be implemented through one or more software. Alternatively, it may be implemented through hardware and software.

The scope of the present invention is not limited to the above-described embodiments, but may be implemented in various forms within the scope of the appended claims. Without departing from the gist of the present invention claimed in the claims, it is considered to be within the scope of the claims of the present invention to various extents that can be modified by any person skilled in the art to which the invention pertains.

1: Fine dust reading system through artificial intelligence-based chroma residual calculation
10: video acquisition unit 20: continuous image extraction unit
30: saturation data calculation unit 40: residual calculation unit
50: fine dust degree calculation unit

Claims (11)

A video acquisition step of acquiring a video indoors or outdoors;
a continuous image extraction step of extracting a plurality of continuous still images from the moving picture;
a chroma data calculation step of calculating a chroma data set for each of the still images;
a residual calculation step of calculating residual data between two consecutive chroma data sets among the plurality of chroma data sets; and
Including; a fine dust degree calculating step of calculating the degree of fine dust based on the residual data;
In the step of calculating the degree of fine dust, the degree of fine dust is calculated by dividing it into classes, and the degree of fine dust is calculated through a pair of influence variables consisting of two variables in which the dominance of the frequency of each other in different classes is reversed. Fine dust reading solution by calculating the residual saturation based on artificial intelligence.
delete delete According to claim 1,
The pair of influence variables is a fine dust reading solution through artificial intelligence-based chroma residual calculation, characterized in that it is determined as a value when the result value of Equation 1 below is maximum.
[Equation 1]

(

: The absolute value of the difference in interest in the class with a low level of fine dust and the level of interest in the class with a high level of fine dust when the two variables are x and y,

: When the two variables are x and y, the degree of interest in the class with a low level of fine dust,

: Interest in a class with a high degree of fine dust when two variables are x and y)
5. The method of claim 4,
Fine through artificial intelligence-based chroma residual calculation, characterized in that when the result value of Equation 1 is the same, two variables in the case where the sum of the frequencies of the two variables in each class is the maximum is determined as the influencing variable pair Dust reading solution.
5. The method of claim 4,
The pair of influence variables is a fine dust reading solution through artificial intelligence-based chroma residual calculation, characterized in that it is derived by learning the residual data through an artificial intelligence analysis model.
According to claim 1,
In the fine dust level calculation step, fine dust reading solution through artificial intelligence-based chroma residual calculation, characterized in that the fine dust level is calculated using an odd number of pairs of influence variables of 3 or more.
According to claim 1,
The fine dust level calculation step is,
A detailed class calculation step of re-determining the degree of fine dust calculated through the influencing variable pair through a second influencing variable pair consisting of two second variables in which the dominance of the frequency in different subclasses is reversed Fine dust reading solution through artificial intelligence-based chroma residual calculation.
According to claim 1,
progressed between the residual calculation step and the fine dust degree calculation step,
Fine dust reading solution through artificial intelligence-based chroma residual calculation, characterized in that it further comprises a residual data filtering step of excluding a component greater than a predetermined value among the components of the residual data.
a video acquisition unit for acquiring a video indoors or outdoors;
a continuous image extraction unit for extracting a plurality of continuous still images from the moving picture;
a chroma data calculator for calculating a chroma data set of each of the still images;
a residual calculator configured to calculate residual data between two consecutive chroma data sets among the plurality of chroma data sets; and
It includes; a fine dust degree calculation unit for calculating the degree of fine dust based on the residual data;
The fine dust level calculation unit calculates the fine dust level by dividing it into classes,
Fine dust reading system through artificial intelligence-based chroma residual calculation, characterized in that the degree of fine dust is calculated using a pair of influence variables consisting of two variables whose frequency dominance is reversed in different classes.
delete

Images