KR102575675B1 - Hyperspectral imaging-based waterworks monitoring and analysis device and method - Google Patents

Abstract

The present invention relates to an apparatus and method for monitoring and analyzing water supply facilities based on hyperspectral imaging. a data analyzer configured to analyze the hyperspectral image and detect a change in the water supply facility; an abnormality detection unit for detecting an abnormality in the waterworks facility when the change in the waterworks facility exceeds a preset reference range; and a notification providing unit generating and providing a notification regarding the detection of the abnormality.

Description

Hyperspectral imaging-based water supply facility monitoring and analysis device and method {HYPERSPECTRAL IMAGING-BASED WATERWORKS MONITORING AND ANALYSIS DEVICE AND METHOD}

The present invention relates to a technology for monitoring facilities, and more particularly, to a hyperspectral imaging-based water supply facility monitoring and analysis device and method capable of performing accurate water supply pipe diagnosis using a data analysis algorithm based on a hyperspectral sensor. will be.

Hyperspectral sensor divides the incident light and acquires tens to hundreds of spectral information in a continuous and narrow wavelength range of the index corresponding to each pixel of the image, so that the unique optical properties and absorption of each material and a sensor capable of analyzing reflection characteristics.

A hyperspectral image obtained from a hyperspectral sensor may correspond to an image having spectral bands of hundreds of consecutive spectral wavelength bands reflected from an object, and information may be processed and collected through an electromagnetic spectrum. Using hyperspectral imaging, it is possible to collect a unique spectrum for each pixel of a desired image by identifying the components of a desired object or detecting a process.

While humans can only see three bands of the visible band of red, green, and blue, hyperspectral sensors can distinguish objects using a wide part of the electromagnetic spectrum, and the range can only be extended beyond the visible band. Rather, the spectrum can be divided into more bands. The spectrum recorded in the hyperspectral image has high wavelength resolution, handles a wide range of wavelength bands, and can include spatial spectral information from materials within a given screen. Therefore, each pixel contains a respective spectral attribute so that the spectra can enable identification of the material constituting the scanned object.

A hyperspectral sensor can obtain information about an object in the form of an image, and each object can appear as a narrow wavelength in a range of the electromagnetic spectrum known as a spectral band. A spectral image can be presented in three dimensions (x, y, λ) consisting of two spatial dimensions and one spectral size. In this case, x and y may represent two spatial dimensions of the scene and λ may represent a spectrum size. Collecting spectral information at each pixel in a two-dimensional (2-D) detector can generate three-dimensional (3-D) data, spatial and spectral information known as a hyperspectral cube. These processes enable the construction of hyperspectral sensors and processing systems in many fields, including astronomy, agriculture, biomedicine, mineralogy and physics.

Korean Publication No. 10-2019-0027419 (2019.03.15)

One embodiment of the present invention is to provide a hyperspectral imaging-based water supply facility monitoring and analysis device and method capable of accurately diagnosing water supply pipes using a data analysis algorithm based on a hyperspectral sensor.

Among embodiments, an apparatus for monitoring and analyzing water supply facilities based on hyperspectral imaging may include a data collection unit configured to collect a hyperspectral image including at least a part of a water supply facility from at least one hyperspectral sensor; a data analyzer configured to analyze the hyperspectral image and detect a change in the water supply facility; an abnormality detection unit for detecting an abnormality in the waterworks facility when the change in the waterworks facility exceeds a preset reference range; and a notification providing unit generating and providing a notification regarding the detection of the abnormality.

The data analysis unit extracts a spectrum of a predetermined band from each of the hyperspectral images continuously extracted at the same first time interval from the hyperspectral image, analyzes the spectrum, and analyzes the region of interest (Region of Interest, ROI) can be determined.

The data analyzer may determine a candidate location for determining the ROI based on a change amount of at least one of an average, a peak, and a gradient of the spectrum.

The data analyzer may determine a region of interest associated with the region of interest from among a plurality of time intervals formed between the hyperspectral images.

The data analysis unit receives the hyperspectral image and the ROI as inputs and generates encoded data related to the ROI, and receives the encoded data as inputs to restore the hyperspectral image related to the ROI. A deep learning model including a decoder is constructed, and feature information on a region of interest of each of the hyperspectral images extracted at the first time interval using the encoder of the deep learning model is used as the encoding data. can create

The data analysis unit extracts successive hyperspectral images at every second time interval smaller than the first time interval within the interest interval from the hyperspectral image, and characteristic information for detecting the anomaly from the interest region of the corresponding hyperspectral images. can create

The anomaly detection unit may predict an anomaly of the water supply facility by inputting the feature information to a pre-constructed anomaly detection model.

The abnormality detection unit may classify the abnormality of the waterworks facility into crack, corrosion, deformation or damage according to the change pattern of the waterworks facility.

Among embodiments, a hyperspectral imaging-based water supply facility monitoring and analysis method may include collecting a hyperspectral image including at least a portion of a water supply facility from at least one hyperspectral sensor through a data collection unit; detecting a change in the waterworks facility by analyzing the hyperspectral image through a data analyzer; detecting an abnormality in the water supply facility through an abnormality detection unit when a change in the water supply facility exceeds a predetermined reference range; and generating and providing a notification regarding the detection of the abnormality through a notification providing unit.

The disclosed technology may have the following effects. However, it does not mean that a specific embodiment must include all of the following effects or only the following effects, so it should not be understood that the scope of rights of the disclosed technology is limited thereby.

Through the hyperspectral imaging-based monitoring and analysis device and method for water supply facilities according to an embodiment of the present invention, it is possible to efficiently operate and maintain water purification plants, sewage treatment plants, water pipe networks, and sewage pipe facilities, and continuously pollute water and sewage through a monitoring system. It is possible to respond early to water quality accidents according to substances.

In addition, through the use of hyperspectral imaging technology, it is possible to efficiently monitor and accurately diagnose the natural environment and water supply and sewerage facilities, which are difficult to evaluate with the naked eye, and enable efficient maintenance in connection with the smart integrated water management system. Accurate judgment based on quantitative data and detailed diagnosis and investigation can be possible by breaking away from qualitative management by

In addition, it is possible to drastically reduce manpower and costs required for monitoring water quality and diagnosing facilities, which are water systems that have a wide range of facilities and require continuous monitoring, and underground facilities that are difficult to directly measure and evaluate. Water resource management and monitoring may be possible.

1 is a diagram illustrating a water supply facility monitoring and analysis system according to the present invention.
FIG. 2 is a diagram explaining the system configuration of the monitoring and analysis device of FIG. 1 .
FIG. 3 is a diagram explaining the functional configuration of the monitoring and analysis device of FIG. 1 .
4 is a flowchart illustrating a method for monitoring and analyzing water supply facilities based on hyperspectral imaging according to the present invention.
5 is a diagram explaining a system configuration of a hyperspectral sensor.
6 is a diagram illustrating a process of collecting information for each pixel of a hyperspectral image through a hyperspectral sensor.

Since the description of the present invention is only an embodiment for structural or functional description, the scope of the present invention should not be construed as being limited by the embodiments described in the text. That is, since the embodiment can be changed in various ways and can have various forms, it should be understood that the scope of the present invention includes equivalents capable of realizing the technical idea. In addition, since the object or effect presented in the present invention does not mean that a specific embodiment should include all of them or only such effects, the scope of the present invention should not be construed as being limited thereto.

Meanwhile, the meaning of terms described in this application should be understood as follows.

Terms such as "first" and "second" are used to distinguish one component from another, and the scope of rights should not be limited by these terms. For example, a first element may be termed a second element, and similarly, a second element may be termed a first element.

It should be understood that when an element is referred to as being “connected” to another element, it may be directly connected to the other element, but other elements may exist in the middle. On the other hand, when an element is referred to as being "directly connected" to another element, it should be understood that no intervening elements exist. Meanwhile, other expressions describing the relationship between components, such as “between” and “immediately between” or “adjacent to” and “directly adjacent to” should be interpreted similarly.

Expressions in the singular number should be understood to include plural expressions unless the context clearly dictates otherwise, and terms such as “comprise” or “having” refer to an embodied feature, number, step, operation, component, part, or these. It should be understood that it is intended to indicate that a combination exists, and does not preclude the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

In each step, the identification code (eg, a, b, c, etc.) is used for convenience of explanation, and the identification code does not describe the order of each step, and each step clearly follows a specific order in context. Unless otherwise specified, it may occur in a different order than specified. That is, each step may occur in the same order as specified, may be performed substantially simultaneously, or may be performed in the reverse order.

The present invention can be implemented as computer readable code on a computer readable recording medium, and the computer readable recording medium includes all types of recording devices storing data that can be read by a computer system. . Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, and optical data storage devices. In addition, the computer-readable recording medium may be distributed to computer systems connected through a network, so that computer-readable codes may be stored and executed in a distributed manner.

All terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs, unless defined otherwise. Terms defined in commonly used dictionaries should be interpreted as consistent with meanings in the context of the related art, and cannot be interpreted as having ideal or excessively formal meanings unless explicitly defined in the present application.

1 is a diagram illustrating a water supply facility monitoring and analysis system according to the present invention.

Referring to FIG. 1 , the water supply facility monitoring and analysis system 100 may include a hyperspectral sensor 110 , a monitoring and analysis device 130 and a database 150 .

The hyperspectral sensor 110 may correspond to a sensor capable of generating a hyperspectral image of an object to be measured, and may be implemented as a computing device capable of storing and transmitting the hyperspectral image to the outside. That is, the hyperspectral sensor 110 may be connected to the monitoring and analysis device 130 through a wired or wireless network, and the plurality of hyperspectral sensors 110 may be simultaneously connected to the monitoring and analysis device 130 . For example, the hyperspectral sensor 110 may be directly coupled to, or installed nearby, water supply facilities such as water purification plants, sewage treatment plants, water pipe networks and sewerage facilities, and may monitor and analyze collected data in real time or periodically. device 130.

The monitoring and analysis device 130 may correspond to a computer or server capable of detecting and reporting an abnormality of water supply facilities by analyzing the hyperspectral image collected through the hyperspectral sensor 110 . The monitoring and analysis device 130 may be connected to the hyperspectral sensor 110 through a network and exchange information. In one embodiment, the monitoring and analysis device 130 may store data necessary for facility monitoring and abnormality detection in association with the database 150 .

The database 150 may correspond to a storage device for storing various pieces of information necessary for the operation of the monitoring and analysis device 130 . The database 150 may store hyperspectral images and images collected during the facility monitoring process, and may store information on algorithms and learning models for data analysis, but is not limited thereto, and the monitoring and analysis device 130 may store In the process of monitoring and analyzing water supply facilities based on hyperspectral imaging, collected or processed information in various forms can be stored.

FIG. 2 is a diagram explaining the system configuration of the monitoring and analysis device of FIG. 1 .

Referring to FIG. 2 , the monitoring and analysis device 130 may include a processor 210 , a memory 230 , a user input/output unit 250 and a network input/output unit 270 .

The processor 210 may execute a water supply facility monitoring and analysis procedure based on hyperspectral imaging according to an embodiment of the present invention, manage the memory 230 read or written in this process, and store the data in the memory 230. You can schedule the synchronization time between volatile memory and non-volatile memory. The processor 210 can control the overall operation of the monitoring and analysis device 130, and is electrically connected to the memory 230, the user input/output unit 250, and the network input/output unit 270 to control data flow between them. can The processor 210 may be implemented as a Central Processing Unit (CPU) or a Graphics Processing Unit (GPU) of the monitoring and analysis device 130 .

The memory 230 may include a secondary storage device implemented as a non-volatile memory such as a solid state disk (SSD) or a hard disk drive (HDD) and used to store all data necessary for the monitoring and analysis device 130, It may include a main memory implemented as a volatile memory such as RAM (Random Access Memory). In addition, the memory 230 may store a set of instructions for executing the hyperspectral imaging-based water supply facility monitoring and analysis method according to the present invention by being executed by the electrically connected processor 210 .

The user input/output unit 250 includes an environment for receiving a user input and an environment for outputting specific information to the user, and includes an adapter such as a touch pad, a touch screen, an on-screen keyboard, or a pointing device. It may include devices and output devices including adapters such as monitors or touch screens. In one embodiment, the user input/output unit 250 may correspond to a computing device connected through a remote connection, and in such a case, the monitoring and analysis device 130 may be implemented as an independent server.

The network input/output unit 270 provides a communication environment for connection with an external device through a network, and includes, for example, a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), and a value VAN (VAN). Added Network) may include an adapter for communication. In addition, the network input/output unit 270 may be implemented to provide a short-range communication function such as WiFi or Bluetooth or a 4G or higher wireless communication function for wireless transmission of data.

FIG. 3 is a diagram explaining the functional configuration of the monitoring and analysis device of FIG. 1 .

Referring to FIG. 3, the monitoring and analysis device 130 includes a data collection unit 310, a data analysis unit 330, an anomaly detection unit 350, a notification providing unit 370, and a control unit (not shown in FIG. 3). can include

At this time, the surveillance and analysis device 130 according to an embodiment of the present invention does not have to include all of the above functional components at the same time, and some of the above components may be omitted or selected from among the above components according to each embodiment. It may be implemented by selectively including some or all of them. Hereinafter, the operation of each component will be described in detail.

The data collection unit 310 may collect a hyperspectral image including at least a part of the water supply facility from the at least one hyperspectral sensor 110 . To this end, the hyperspectral sensor 110 may be directly coupled to a water supply facility or may be installed and operated nearby, and may be connected to the monitoring and analysis device 130 through a network. For example, when the water supply facility is a water supply pipe, a plurality of hyperspectral sensors 110 may be installed at predetermined intervals along the water supply pipe, and a hyperspectral image is obtained for each section of the water supply pipe by the hyperspectral sensors 110. this can be collected.

In addition, the hyperspectral sensor 110 may be implemented in a form portable by a person, such as a portable hyperspectral camera or a hyperspectral camera for drones. Alternatively, hyperspectral images directly measured by the user for the suspicious section may be collected.

In an embodiment, the data collection unit 310 may collect images of water supply facilities through various image sensors such as a multi-spectral camera, a thermal imaging camera, and an endoscope camera instead of the hyperspectral sensor 110 . Meanwhile, the data collection unit 310 is not limited to the above methods and can collect images and data similar to the hyperspectral image through a measuring device similar to the hyperspectral sensor 110 .

In addition, the data collection unit 310 may store the hyperspectral image collected from the hyperspectral sensor 110 in the database 150 and may perform a preprocessing operation as needed. For example, the data collection unit 310 may perform preprocessing operations such as image segmentation, rotation, and correction.

The data analyzer 330 may analyze the hyperspectral image to detect changes in water supply facilities. The data analyzer 330 may analyze the hyperspectral images in chronological order, and may monitor changes in facilities through a difference between hyperspectral images continuously collected through the same hyperspectral sensor 110 . For example, if a difference in a specific area exceeds a normal range in hyperspectral images collected consecutively, the data analysis unit 330 determines that a change has occurred in the waterworks facility corresponding to the corresponding area and performs a more precise analysis operation. This can be initiated.

In one embodiment, the data analysis unit 330 extracts a spectrum of a predetermined band from each of the hyperspectral images continuously extracted at the same first time interval from the hyperspectral image and analyzes the spectrum to analyze the region of interest for the water supply facility. (Region of Interest, ROI) can be determined. The data analyzer 330 may generate a hyperspectral image divided by frame from the hyperspectral image, and may extract the hyperspectral image at the same time interval.

For example, the data analyzer 330 may generate a hyperspectral image at regular intervals by applying time intervals such as seconds, minutes, and hours. The data analyzer 330 may extract a spectrum of a visible ray band or an infrared band from each of the hyperspectral images, and may extract a spectrum of both the visible ray band and the infrared band, if necessary. Meanwhile, the data analyzer 330 may extract the spectrum from the hyperspectral image through various methods, and description of the specific method is omitted.

In addition, the data analyzer 330 may determine a region having a relatively large difference as a region of interest through comparison between temporally consecutive images among the extracted hyperspectral images. To this end, the data analyzer 330 may collect pixels with a large difference through comparison between the spectra of each image, and focus on a region with a relatively large difference and existing in an adjacent position through clustering based on pixel positions. area can be determined. That is, the region of interest may correspond to a portion where a large change occurs for a certain period of time in a water supply facility.

In an embodiment, the data analyzer 330 may determine a candidate position for determining a region of interest based on a change amount of at least one of a mean, a peak, and a gradient of a spectrum. The data analyzer 330 may extract a spectrum from the hyperspectral image and then calculate information such as an average, peak, and slope of the spectrum. The data analyzer 330 may determine, as candidate positions of the ROI, pixels having a variation amount of spectral feature values exceeding a preset threshold value. That is, the data analyzer 330 may determine a region of interest based on temporal or regional similarities between candidate locations. Meanwhile, the data analyzer 330 may normalize the spectrum and determine a region of interest using the normalized spectrum. Also, the data analyzer 330 may determine a region of interest based on a change in reflectance or absorption of the spectrum.

In one embodiment, the data analysis unit 330 may determine a region of interest associated with the region of interest from among a plurality of time intervals formed between hyperspectral images. Changes in water supply facilities may be temporary, and the data analysis unit 330 may extract hyperspectral images during a time interval at a point in time when a change occurs in the hyperspectral image based on the region of interest. In particular, when the hyperspectral images are extracted at regular time intervals, a section most temporally related to a change in water supply facilities among time sections between the hyperspectral images may be determined as a section of interest. For example, a time interval in which the difference between the spectra of hyperspectral images is the largest may be determined as the interest interval, and a time interval in which a change in the mean, peak, or slope of the spectrum is greatest may be determined as the interest interval, but is not necessarily limited thereto. Of course not.

In one embodiment, the data analysis unit 330 may build an encoder-decoder-based deep learning model and perform analysis on the hyperspectral image using the deep learning model. Specifically, the deep learning model may include an encoder and a decoder. For example, the deep learning model may be implemented through an autoencoder model, but is not necessarily limited thereto. The encoder may convert the hyperspectral image into encoded data, and the decoder may reconstruct the encoded data into a hyperspectral image. In particular, an encoder may be defined to receive a hyperspectral image and a region of interest as inputs to generate encoded data relating to a region of interest, and a decoder may be defined to receive encoded data as inputs and reconstruct a hyperspectral image associated with a region of interest. can As a result, the data analyzer 330 may generate feature information on the ROI of each of the hyperspectral images extracted at the first time interval as encoding data by using the encoder of the deep learning model.

In one embodiment, the data analysis unit 330 extracts successive hyperspectral images at every second time interval smaller than the first time interval within the ROI from the hyperspectral image, and detects an anomaly from the ROI of the corresponding hyperspectral images. It is possible to generate feature information for. The data analyzer 330 may apply a second time interval corresponding to a collection interval smaller than the first time interval in order to collect more hyperspectral images for the section of interest of the hyperspectral image. Accordingly, the data analyzer 330 may effectively analyze changes in water supply facilities through comparison between regions of interest in hyperspectral images continuously collected at shorter intervals.

The anomaly detection unit 350 may detect an anomaly in a waterworks facility when a change in the waterworks facility exceeds a preset reference range. At this time, the change of the water supply facility may be determined based on the difference between the hyperspectral images. That is, the anomaly detection unit 350 may determine that the change in the water supply facility has occurred abnormally as the difference between the hyperspectral images increases, and may set and utilize a reference range for anomaly detection in advance.

In one embodiment, the anomaly detection unit 350 may predict an anomaly of a water supply facility by inputting feature information to a pre-constructed anomaly detection model. Here, the anomaly detection model may correspond to a deep learning model that receives feature information about a region of interest as an input and generates predictive information about anomaly of a water supply facility as an output. The anomaly detection unit 350 provides feature information on the region of interest generated by the data analysis unit 330 as an input to an anomaly detection model to receive probability information about an anomaly in the water supply facility, and based on the probability information, the water supply facility abnormalities can be detected.

In one embodiment, the abnormality detection unit 350 may classify the abnormality of the waterworks facility into crack, corrosion, deformation, or damage according to the change pattern of the waterworks facility. The anomaly detection unit 350 may obtain time-series information on changes in water supply facilities through comparison between hyperspectral images, and may generate a change pattern through the time-series information. The anomaly detection unit 350 may generate specific anomaly information by comparing a change pattern of water supply facilities with previously established pattern data. Cracks may occur due to structural defects or defects in facilities, corrosion may occur due to various electrochemical reactions of metallic materials, deformation may occur due to the load or pressure of facilities, and damage may occur due to deterioration and deterioration of facilities. It can be caused by an external shock, etc. The anomaly detection unit 350 may map anomalies of waterworks facilities to various change patterns and define them in advance, and may effectively predict anomalies of waterworks facilities by classifying analysis results of hyperspectral images.

The notification providing unit 370 may generate and provide a notification related to abnormality detection. The notification providing unit 370 may provide notification information to terminals registered in advance, such as users, managers, and facility managers, when an abnormality is detected by the abnormality detection unit 350 . In this case, the notification information may be implemented and provided in various ways such as e-mail, text message, push message, sound, and vibration.

The control unit (not shown in FIG. 3) controls the overall operation of the monitoring and analysis device 130, and the communication between the data collection unit 310, the data analysis unit 330, the anomaly detection unit 350, and the notification providing unit 370 It can manage control flow or data flow.

4 is a flowchart illustrating a method for monitoring and analyzing water supply facilities based on hyperspectral imaging according to the present invention.

Referring to FIG. 4 , the monitoring and analysis device 130 may collect a hyperspectral image including at least a part of the water supply facility from at least one hyperspectral sensor through the data collection unit 310 (step S410). The monitoring and analysis device 130 may analyze the hyperspectral image through the data analysis unit 330 to detect changes in water supply facilities (step S430).

In addition, the monitoring and analysis device 130 may detect an abnormality in the water supply facility through the abnormality detection unit 350 when the change in the water supply facility exceeds a predetermined standard range (step S450). The monitoring and analysis device 130 may generate and provide a notification regarding abnormality detection through the notification providing unit 370 (step S470).

5 is a diagram explaining a system configuration of a hyperspectral sensor.

Referring to FIG. 5 , the system configuration of the hyperspectral sensor can be largely divided into three areas. Specifically, the hyperspectral sensor may include a front optical unit (A) that scans and measures an object, a spectrometer (B) that splits received light, and a detector (C) that converts the split light into a spectrum.

The front optical unit may include an optical lens. An optical lens may include a focal length and an aperture value (F value). The focal length may correspond to the distance from the main point of the optics to the focal point, and the shorter the focal length, the wider FOV can be formed. The diaphragm can control the amount of light passing through the lens, and the diaphragm value may correspond to numerical values of the degree of opening of the lens. That is, the lower the aperture value, the more the aperture can be opened.

The wavelength resolution capability and sensitivity of the spectrometer may be determined according to the shape of the dispersing element. The detector may convert light into an electrical signal through an image sensor, and the image sensor may mainly use CCD, CMOS, or the like depending on a manufacturing process and application method for generating an image into a digital signal.

6 is a diagram illustrating a process of collecting information for each pixel of a hyperspectral image through a hyperspectral sensor.

Referring to FIG. 6 , the monitoring and analysis device 130 may collect hyperspectral images of water supply facilities through a hyperspectral sensor. The monitoring and analysis device 130 may analyze the hyperspectral image to detect changes in water supply facilities. At this time, the monitoring and analysis device 130 may generate feature information about the region of interest of the water supply facility by using spectral information collected at the same location in each of the hyperspectral images of the hyperspectral image. The monitoring and analysis device 130 can provide efficient and accurate management and monitoring while dramatically reducing manpower and costs required for monitoring and diagnosing facilities by detecting abnormalities in water supply facilities and promptly notifying them using feature information.

Although the above has been described with reference to preferred embodiments of the present invention, those skilled in the art will variously modify and change the present invention within the scope not departing from the spirit and scope of the present invention described in the claims below. You will understand that it can be done.

100: water supply facility monitoring and analysis system
110: hyperspectral sensor 130: monitoring analysis device
150: database
210: processor 230: memory
250: user input/output unit 270: network input/output unit
310: data collection unit 330: data analysis unit
350: Abnormality detection unit 370: Notification providing unit

Claims (9)

a data collection unit that collects a hyperspectral image including at least a portion of a water supply facility from at least one hyperspectral sensor;
Analyzing the hyperspectral image to detect changes in the water supply facility, extracting a spectrum of a preset band from each of the hyperspectral images continuously extracted at the same first time interval from the hyperspectral image, and analyzing the spectrum a data analysis unit determining a region of interest (ROI) for the water supply facility;
an abnormality detection unit for detecting an abnormality in the waterworks facility when the change in the waterworks facility exceeds a preset reference range; and
A notification providing unit generating and providing a notification related to the detection of the abnormality;
The data analysis unit receives the hyperspectral image and the ROI as inputs and generates encoded data related to the ROI, and receives the encoded data as inputs to restore the hyperspectral image related to the ROI. A deep learning model based on an encoder-decoder including a decoder is constructed, and feature information on a region of interest of each of the hyperspectral images extracted at the first time interval using the encoder of the deep learning model is used. as the encoding data,
The hyperspectral imaging-based water supply facility monitoring and analysis device, characterized in that the anomaly detection unit predicts an anomaly of the water supply facility by inputting the feature information into a pre-constructed anomaly detection model.
delete The method of claim 1, wherein the data analysis unit
A hyperspectral imaging-based water supply facility monitoring and analysis device, characterized in that for determining a candidate position for determining the region of interest based on a change in at least one of the average, peak, and gradient of the spectrum.
The method of claim 1, wherein the data analysis unit
A hyperspectral imaging-based water supply facility monitoring and analysis device, characterized in that for determining a section of interest associated with the region of interest from among a plurality of time sections formed between the hyperspectral images.
delete The method of claim 4, wherein the data analysis unit
Extracting successive hyperspectral images at every second time interval smaller than the first time interval within the interest interval from the hyperspectral image and generating feature information for detecting the anomaly from the region of interest of the corresponding hyperspectral images. Hyperspectral imaging-based monitoring and analysis device for water supply facilities.
delete The method of claim 1, wherein the abnormality detection unit
A hyperspectral imaging-based water supply facility monitoring and analysis device, characterized in that the abnormality of the water supply facility is classified as crack, corrosion, deformation or damage according to the change pattern of the water supply facility.
In the water supply facility monitoring and analysis method performed in the water supply facility monitoring and analysis device,
Collecting a hyperspectral image including at least a part of a waterworks facility from at least one hyperspectral sensor through a data collection unit;
Through the data analysis unit, the hyperspectral image is analyzed to detect a change in the water supply facility, and a spectrum of a predetermined band is extracted from each of the hyperspectral images continuously extracted at the same first time interval from the hyperspectral image, analyzing the spectrum to determine a region of interest (ROI) for the water supply facility;
detecting an abnormality in the water supply facility through an abnormality detection unit when a change in the water supply facility exceeds a predetermined reference range; and
Generating and providing a notification related to the detection of the abnormality through a notification providing unit; including,
The determining of the ROI may include: an encoder generating encoded data related to the ROI by receiving the hyperspectral image and the ROI as inputs; and receiving the encoded data as inputs to generate encoding data related to the ROI. An encoder-decoder-based deep learning model including a decoder for reconstructing a spectral image is constructed, and a region of interest of each of the hyperspectral images extracted at the first time interval using the encoder of the deep learning model. Generating feature information for as the encoded data;
Wherein the step of detecting the anomaly comprises predicting the anomaly of the water supply facility by inputting feature information into a pre-constructed anomaly detection model.