KR102529265B1 - Apparatus for sensing leakage using image and sound and method thereof - Google Patents

Abstract

According to an embodiment of the present invention, based on a video signal and a sound signal, an image change area in which an image change occurs is primarily detected, and a sound signal generated by leakage is secondarily detected in the image change area to more accurately determine the authenticity of a leak. It is to provide a leak detection device and method using images and sounds that can be used. A leak detection device using image and sound according to an embodiment of the present invention includes an image unit for acquiring a video signal for a leak monitoring object, an audio unit for acquiring an audio signal for a leak monitoring object, and an image unit and an audio unit. and a control unit that firstly analyzes an image change region in which an image change occurs over time using the acquired image signal and sound signal, and secondarily analyzes a sound signal generated by leakage in the image change region.

Description

Leak detection device and method using image and sound {APPARATUS FOR SENSING LEAKAGE USING IMAGE AND SOUND AND METHOD THEREOF}

The present invention relates to a leak detection device and method using images and sounds.

In domestic nuclear power plants, a leak monitoring system is installed and operated to detect leaks in the reactor system (safety system composed of pressure vessel and piping) at an early stage. In such a leak monitoring system, acoustic emission (AE) sensors for detecting elastic waves generated when a reactor system leaks are distributed and installed on the surface of the system. In the leak monitoring system, when the AE sensor detects an elastic wave generated at the leak part of the system surface, the signal measured by the AE sensors is analyzed to estimate the location where the leak occurs. At this time, since the elastic wave propagated along the surface of the system structure due to leakage has a damping characteristic, it is not easy to accurately determine the authenticity of the leakage when the distance from the leakage occurrence location to the AE sensor is long. In addition, when leakage occurs in a system (area) where an AE sensor is not installed, there is a disadvantage in that leakage monitoring is not easy.

As a related prior art document, Japanese Patent Registration No. 3,295,682 discloses a "signal processing method for specifying a pipe leakage location".

Japanese registered patent 3,295,682

According to an embodiment of the present invention, based on a video signal and a sound signal, an image change area in which an image change occurs is primarily detected, and a sound signal generated by leakage is secondarily detected in the image change area to more accurately determine the authenticity of a leak. It is to provide a leak detection device and method using images and sounds that can be used.

In addition to the above tasks, embodiments according to the present invention may be used to achieve other tasks not specifically mentioned.

A leak detection device using image and sound according to an embodiment of the present invention includes an image unit for acquiring a video signal for a leak monitoring object, an audio unit for acquiring an audio signal for a leak monitoring object, and an image unit and an audio unit. and a control unit that firstly analyzes an image change region in which an image change occurs over time using the acquired image signal and sound signal, and secondarily analyzes a sound signal generated by leakage in the image change region.

)을 산출하는 단계, 누설 후보영역(

)을 산출하는 단계, 그리고 시간에 따라 변화하는 영상변화 영역에서 누설에 의한 음향이 발생하는지를 판별하고, 누설에 의한 음향 발생이면 누설발생을 경보하는 단계를 포함한다.On the other hand, a leak detection method using video and sound according to an embodiment of the present invention includes the steps of acquiring a video signal and an audio signal of an object to be monitored for leakage, and using the acquired video signal to detect each image change area that changes over time. area representative value (

), the leakage candidate region (

), and determining whether sound due to leakage is generated in the image change area that changes with time, and alerting the occurrence of leakage if sound is generated due to leakage.

An embodiment of the present invention is to improve the leakage monitoring function for pressure vessels and piping by additionally determining whether an acoustic signal due to leakage is generated in the image change area where the image change occurs by using a video signal and an audio signal together. There are possible effects.

1 is a diagram illustrating a leak detection device using video and sound according to an embodiment of the present invention.
2 is a flowchart illustrating a leak detection method using video and sound according to an embodiment of the present invention.
3 is a flowchart showing a leakage detection state using video and sound according to an embodiment of the present invention.

With reference to the accompanying drawings, embodiments of the present invention will be described in detail so that those skilled in the art can easily carry out the present invention. This invention may be embodied in many different forms and is not limited to the embodiments set forth herein. In order to clearly describe the present invention in the drawings, parts irrelevant to the description are omitted, and the same reference numerals are used for the same or similar components throughout the specification. In addition, in the case of widely known known technologies, detailed descriptions thereof will be omitted.

Throughout the specification, when a certain component is said to "include", it means that it may further include other components without excluding other components unless otherwise stated.

Hereinafter, a leak detection device and method using video and sound will be described in detail with reference to the drawings.

1 is a diagram illustrating a leak detection device using video and sound according to an embodiment of the present invention, and FIG. 2 is a flowchart illustrating a leak detection method using video and sound according to an embodiment of the present invention. And Figure 3 is a flow chart showing a leak detection state using video and sound according to an embodiment of the present invention. Referring to FIG. 1, the leak detection device using image and sound includes an image unit 100, a sound unit 110, a control unit 120, a display unit 130, and an alarm unit 140, and includes a video signal and a sound signal. Based on this, it is possible to easily determine the authenticity or counterfeit of the leak monitoring object by firstly detecting the image change area where the image change occurs and by secondarily detecting the generation of an acoustic signal due to leakage in the image change area. Here, the leak monitoring object may include a high/low temperature pressure vessel and a pipe.

The imaging unit 100 may include an image detection unit that acquires an image signal for an object to be monitored for leakage. The imaging unit 100 may include a cam or CCTV for image acquisition.

The sound unit 110 may include a sound detector that acquires a sound signal for an object to be monitored for leakage. The sound unit 110 may include three or more microphones. A plurality of sound units 110 may be installed at intervals from each other in the area where the leak monitoring object is located.

The control unit 120 first analyzes the image change region where the image change occurs over time using the image signal and the sound signal acquired from the image unit 100 and the sound unit 110, and then Acoustic signal generation can be analyzed secondarily. The image signal and the sound signal respectively acquired through the image unit 100 and the sound unit 110 may be acquired simultaneously. The control unit 120 may include a microprocessor that performs overall control operations related to detecting and processing the leak location of the leak monitoring object. The controller 120 may be connected to a general computer to share related information. The control unit 120 may have a separate memory unit. The memory unit may store information calculated by the control unit 120 and related data for detecting and controlling the leak location of the leak monitoring object.

Meanwhile, the control unit 120 may generate a leak location display signal by detecting a leak component of the object to be monitored for leak. The display unit 130 may display the leak location of the object to be monitored for leak based on the leak location display signal. Also, the control unit 120 may detect a leak component of the object to be monitored for leakage and generate a leak location warning signal. The alarm unit 140 may alert the leak location of the object to be monitored for leakage based on the leak location alarm signal.

As described above, it is possible to determine the authenticity of a leakage monitoring object based on video and audio signals using the leak detection device using video and sound according to an embodiment of the present invention. That is, by utilizing sound signals in addition to video signals to additionally determine whether or not sound signals due to leaks occur in the video change area where image changes occur, monitoring leaks in pressure vessels and pipes and authenticity discrimination functions compared to existing methods. can improve

A detection method using a leak detection device using video and sound according to an embodiment of the present invention will be described with reference to FIGS. 1 to 3 .

)와 음향신호(

)를 동시에 취득할 수 있다. 이때 t는 시간을 나타낸다.First, a sound signal and a video signal are obtained for a leakage monitoring target by using the image unit 100 and the sound unit 110 (S210). Here, the image signal for the leakage monitoring target using the image unit 100 and the sound unit 110 (

) and sound signals (

) can be obtained simultaneously. Here, t represents time.

)을 산출한다(S220). 고온/저온 압력 용기 및 배관에서 누설이 발생하는 경우, 온도차에 의하여 누설부위에서 증기나 가스형태의 누설이 발생할 수 있다. 이러한 누설이 발생하는 영역의 영상은 시간(t)에 따라 변화하므로, 수학식 1 내지 수학식 3을 이용하여 측정된 영상신호 (

)중 시간에 따라 변화하는 영상변화 영역을 확인할 수 있다.Using the acquired video signal, each region representative value of the image change region that changes with time in the image signal (

) is calculated (S220). When leakage occurs in high/low temperature pressure vessels and pipes, leakage in the form of steam or gas may occur at the leakage site due to the temperature difference. Since the image of the region where such leakage occurs changes with time t, the image signal measured using Equations 1 to 3

), it is possible to check the image change area that changes with time.

))을 산출한다(S220).First, as shown in Equation 1, a representative value of each region of an image captured by applying convolution to an image signal (

)) is calculated (S220).

[Equation 1]

: 2차원 합성곱,

; 2차원 합성곱에 적용되는 2차원 함수, 가우시안 함수 등 사용 가능,

와

: 촬영한 영상의 m번째 영역(m=1~M)의 좌표)(here,

: two-dimensional convolution,

; 2D function applied to 2D convolution, Gaussian function, etc. can be used,

and

: Coordinates of the mth area (m=1~M) of the captured image)

)을 산출하는 단계(S230)에서 각 영역의 대표값(

)에 대해 시간변화에 대한 각 영역의 대표값(

)의 변화값(

)을 수학식 3을 이용하여 산출하고 산출된 변화값이 수학식 4와 같이 설정값을 초과하는 경우 누설 후보영역(

)으로 판별할 수 있다.Leakage candidate area (

) In the step of calculating (S230), the representative value of each region (

), the representative value of each area for time change (

) of the change value (

) is calculated using Equation 3, and when the calculated change value exceeds the set value as shown in Equation 4, the leakage candidate area (

) can be identified.

[Equation 2]

[Equation 3]

[Equation 4]

, l=1~L)에 대응하는 영상 위치에 해당하는 사물을 인식하는 방법으로는 사물인식에 사용되는 컨볼루션 신경망(CNN ; Convolutional Neural Network)기법 등을 적용하여, 누설후보영역에 해당하는 사물이 안전계통과 같이 누설경보발생 대상물 인지를 판별하는 방법을 사용할 수 있다. 또한 안전계통과 같이 누설경보발생 대상물이 다른 구조물과 다른 색으로 구분되어 있는 경우에는 누설후보영역에 대응하는 영상 위치에 해당하는 사물이 누설경보발생 대상물에 해당하는 색상인지를 판별하는 방법을 사용할 수 있다.After the leakage candidate area calculation step, a leakage alarm generating object confirmation step (S240) of recognizing an object corresponding to an image area having a large change over time and determining whether the object is a safety system (a leak alarm generating object) may be further included. can Here, the leakage candidate region (

, l = 1 ~ L), as a method of recognizing objects corresponding to image positions corresponding to object recognition, convolutional neural network (CNN) techniques used in object recognition are applied to objects corresponding to the leakage candidate area. Like this safety system, it is possible to use a method for determining whether the leak alarm is an object or not. In addition, if the leak alarm target is distinguished by a different color from other structures, such as a safety system, a method of determining whether the object corresponding to the image position corresponding to the leak candidate area is the color corresponding to the leak alarm target can be used. there is.

)를 이용하여 상호상관함수(

)를 산출하는 단계, 산출된 상호상관함수에 대한 시간지연-주파수 분석을 수행하여, 산출된 상호상관함수의 시간지연-주파수분석 결과(

)를 산출하는 단계, 그리고 산출된 시간지연-주파수분석 결과(

)에서 시간에 따라 변화가 큰 누설 후보영역(

)에 해당하는 시간지연(

와

)에 수직선형태로 표출되는 누설신호가 존재하는 경우 누설에 의한 음향 발생으로 판별하는 단계를 포함할 수 있다. 즉, 시간에 따라 변화가 큰 누설 후보영역(

, l=1~L)에서 누설에 의한 음향이 발생하는지를 판별하는 방법으로는 음향신호(

)에 대한 시간지연-주파수분석(Time delay-Frequency Analysis) 결과(

)에서 시간에 따라 변화가 큰 누설 후보영역(

)에 해당하는 시간지연(

와

)에 수직선형태로 표출되는 누설특징이 존재하는 경우 누설에 의한 음향 발생으로 판별할 수 있다. 이때

는 마이크로폰에서 측정된 음향신호(

,

)를 이용하여 산출된 상호상관함수(

)에 대한 시간지연-주파수분석 결과이다. 이때 3개의 마이크로폰을 한글자음"ㄴ"자 형태로 배치한 경우,

와

이며, C는 음속을 나타내고, d는 마이크로폰 사이의 거리를 나타낸다.Subsequently, the control unit determines whether sound is generated due to leakage in the image area where the change is large over time, and generates an alarm if it is determined as leakage (S250). Here, the determination of sound generation by leakage is a sound signal (

) using the cross-correlation function (

), performing time-delay-frequency analysis on the calculated cross-correlation function, resulting in time-delay-frequency analysis of the calculated cross-correlation function (

), and the calculated time delay-frequency analysis result (

), a leakage candidate region with a large change over time (

) corresponding to the time delay (

and

), if there is a leakage signal expressed in the form of a vertical line, a step of determining that sound is generated due to leakage may be included. That is, the leakage candidate region with a large change over time (

, l = 1 ~ L), as a method of determining whether sound is generated by leakage, the sound signal (

) for the time delay-frequency analysis (Time delay-Frequency Analysis) result (

), a leakage candidate region with a large change over time (

) corresponding to the time delay (

and

), if there is a leakage characteristic expressed in the form of a vertical line, it can be determined as sound generation by leakage. At this time

is the acoustic signal measured by the microphone (

,

), the cross-correlation function calculated using (

) is the result of time delay-frequency analysis. At this time, if the three microphones are arranged in the shape of a Korean consonant "ㄴ",

and

, C represents the speed of sound, and d represents the distance between the microphones.

)는 수학식 5를 이용하여 산출할 수 있다.The cross-correlation function (

) can be calculated using Equation 5.

[Equation 5]

: 기대값,

: 시간지연)(here,

: expected value,

: time delay)

For leak authenticity determination, the user can visually check whether there is a component corresponding to the leakage signal in the time delay-frequency conversion result. For automated determination, image recognition techniques such as convolutional neural network (CNN) techniques Using this method, it is possible to use a method of discriminating through deep learning learning for time delay-frequency analysis images.

As described above, the leakage detection method using video and sound according to an embodiment of the present invention uses only the video signal to detect leakage by using both the video signal and the sound signal information, compared to the technique of monitoring leakage of the object to be monitored. You can accurately determine whether it is or not. Therefore, it is possible to improve the reliability of determining whether or not a leak has occurred in the object to be monitored for leakage.

In the leakage monitoring method using video signals, when a leak occurs in a high or low temperature pressure container or pipe, the image near the leak point among the video signals is changed (difference between It is possible to determine whether there is leakage by analyzing the video signal using the video). Using video signals, leaks can be monitored from a distance in a non-contact manner. However, the leakage monitoring method using a video signal has a disadvantage in that structure vibration occurs or the movement of other objects or the movement of a worker can be determined as leakage. That is, in the case of monitoring the leakage of pressure vessels and pipes by detecting only the change (difference) of the image signal between the case of no leakage and the case of leakage, whether the image change due to leakage of pressure vessels and pipes, vibration of the structure or It is not possible to determine whether the image change is caused by the motion of another object. Therefore, a more accurate leakage monitoring technique is required.

An apparatus and method for detecting leakage using video and sound according to an embodiment of the present invention use an audio signal in addition to a video signal to additionally determine whether an audio signal due to leakage is generated in an image area where an image change occurs, thereby using an existing method. Compared to this, it is possible to improve the leakage monitoring and authenticity determination function for pressure vessels and pipes.

Although the preferred embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements of those skilled in the art using the basic concept of the present invention defined in the following claims are also made according to the present invention. falls within the scope of the rights of

100; video unit 110; sound department
120; control unit 130; display part
140; warning department

Claims (10)

An image unit for acquiring an image signal for a leak monitoring object;
A sound unit for acquiring an acoustic signal for the leak monitoring object, and
A control unit that firstly analyzes an image change region in which an image change occurs over time using the image signal and sound signal obtained from the image unit and the sound unit, and secondly analyzes the generation of a sound signal due to leakage in the image change region.
Including,
The control unit
Each region representative value of the image change region that changes over time using the acquired image signal (

) is the following Equation 1:
[Equation 1]

(here,

: two-dimensional convolution,

; 2D function applied to 2D convolution, Gaussian function, etc. can be used,

and

: Coordinates of the mth area (m=1~M) of the captured image)
calculated using
Leakage candidate area (

) to the representative value of each area (

), the representative value of each area for time change (

) of the change value (

) to Equation 3:
[Equation 3]

(Where, μt[R(xm,ym,t)]: Representative value of each region for time change (

) average value)
If calculated using and the calculated change value exceeds the set value, the leakage candidate area (

), a leak detection device using video and sound. In paragraph 1,
The control unit detects a leakage component of the leak monitoring object to generate a leak location display signal, and further includes a display unit for displaying a leak location of the leak monitoring object based on the leak location display signal. sensor. In paragraph 1,
The control unit detects a leakage component of the leak monitoring object to generate a leak location alarm signal, and further includes an alarm unit for alerting the leak location of the leak monitoring object based on the leak location alarm signal, leak using image and sound. sensor. Acquiring a video signal and an audio signal of a leak monitoring object;
Each region representative value of the image change region that changes over time using the acquired image signal (

) is the following Equation 1:
[Equation 1]

(here,

: two-dimensional convolution,

; 2D function applied to 2D convolution, Gaussian function, etc. can be used,

and

: Calculating using the coordinates of the m-th region (m = 1 to M) of the captured image,
Leakage candidate area (

), and
Determining whether sound due to leakage is generated in the image change area that changes with time, and alarming the occurrence of leakage if sound is generated due to leakage
Including,
The leakage candidate region (

), the representative value of each area (

), the representative value of each area for time change (

) of the change value (

) to Equation 3:
[Equation 3]

(Where, μt[R(xm,ym,t)]: Representative value of each region for time change (

) average value)
If calculated using and the calculated change value exceeds the set value, the leakage candidate area (

), and a leak detection method using video and sound.
delete delete delete In paragraph 4,
Determination of sound generation by the leakage
The sound signal (

) using the cross-correlation function (

) to Equation 5:
[Equation 5]

(here,

: expected value,

: time delay)
The step of calculating using
By performing time-delay-frequency analysis on the calculated cross-correlation function, the result of time-delay-frequency analysis of the calculated cross-correlation function (

), and
Calculated time delay-frequency analysis result (

), a leakage candidate region with a large change over time (

) corresponding to the time delay (

and

), a leak detection method using image and sound, including the step of determining that a leak signal is generated by a leak when there is a leak signal expressed in the form of a vertical line.
delete In paragraph 4,
The leakage detection method using video and sound further comprising a step of checking a leak alarm generating object for determining whether an object corresponding to a change area over time is a safety system after the step of calculating the leak candidate area.

Images