KR102028823B1 - Apparatus and method for detecting uwb-based oil film - Google Patents

Abstract

A UWB based oil film detection apparatus and method are presented. According to one or more exemplary embodiments, an UWB-based oil film detecting apparatus may include: a radar transceiver configured to transmit a signal toward a detection target surface using an ultra wide band (UWB) and receive reflected waves of the transmitted signal to collect data; And an oil film detector for analyzing the data collected by the radar transceiver to detect an oil film on the detection target surface.

Description

UWB-based oil film detection apparatus and method {APPARATUS AND METHOD FOR DETECTING UWB-BASED OIL FILM}

The following embodiments relate to an apparatus and method for detecting an ultra wide band (UWB) based oil film, and more particularly, to an apparatus and method for detecting an oil film using a radar reflection characteristic of water of the UWB. .

Water pollution caused by oil spills, such as oceans and streams, takes a lot of time and effort to recover, as well as a wide range of damages such as ecosystem destruction and property damage. Therefore, in order to minimize oil damage by detecting oil at an early stage, many oil film detectors capable of detecting oil in rivers or purification facilities are installed.

However, expensive oil film detectors use infrared sensors, lasers, microwaves, etc. in a local range, and SAR (Synthetic Aperture Radar), microwave scattering systems, etc. are used as a wide range of detectors such as sea. Among these, small oil film detectors used in small areas such as ships and rivers are directly installed on the surface of the water according to the type of sensor and the detection method.They can measure the oil film and the non-contact method to measure the oil film at a certain height. Contactless applications are often applied because periodic maintenance is required.

A typical non-contact oil film detector using a laser is a technology that determines the refractive index that varies depending on the type and existence of the oil film by irradiating the laser onto the surface of the water, and detects the local area rather than measuring the entire channel. In order to increase the installation position and the height of the surface is also important. Commercially available laser type oil film detectors have a fixed type and many malfunctions occur due to insufficient consideration of the influence of detection performance due to environmental factors such as sleep conditions and fog disturbance.

Korean Laid-Open Patent Publication No. 10-2014-0140039 describes a technique related to an oil film detecting apparatus capable of detecting an oil film even when a distance from such a water surface (detection target surface) is apart.

Korean Patent Publication No. 10-2014-0140039

Embodiments describe a UWB-based oil film detecting apparatus and method, and more specifically, obtains not only the oil film but also the distance from the water surface and the thickness of the oil film by using an ultra wide band (UWB) based on an impulse radio (IR). Provides a UWB-type oil film detection technology that can be.

Embodiments provide a UWB-based oil film detecting apparatus and method for detecting oil film based on deep learning based on distance and type of oil film in order to maintain the distance between the surface and the equipment and to secure the oil film detection accuracy in a non-contact manner. It is.

According to one or more exemplary embodiments, an UWB-based oil film detecting apparatus may include: a radar transceiver configured to transmit a signal toward a detection target surface using an ultra wide band (UWB) and receive reflected waves of the transmitted signal to collect data; And an oil film detector for analyzing the data collected by the radar transceiver to detect an oil film on the detection target surface.

The radar transceiver may detect an oil film using a radar reflection characteristic of water as the signal is transmitted toward the surface of the detection target surface using the UWB in a non-contact manner.

The radar transceiver may acquire information about the presence or absence of the oil film, the type of the oil film, the thickness of the oil film, and the distance to the water surface, which is the detection target surface, using the IR radio based UWB.

The radar transceiver may use a directional antenna or an omnidirectional antenna as necessary to adjust the measurement accuracy or height of the localized portion.

The radar transmitting and receiving unit is configured in a mobile device capable of three-axis adjustment for changing the height or direction of the surface of the surface to be detected, it is possible to automatically adjust the position according to the height change of the surface for optimal transmission and reception of the UWB Do.

The radar transceiver may use a single antenna method capable of transmitting and receiving signals simultaneously.

The oil film detection unit, in order to maintain the distance between the radar transceiver and the water surface being the detection target surface in a non-contact manner and to ensure the detection accuracy of the oil film deep learning (deep learning) capable of classifying the type of oil film The oil film can be detected based on this.

The deep learning may analyze the data collected through a deep learning based oil film detector model using a UWB radar composed of a one-dimensional convolutional neural network (CNN).

The oil film detection unit processes the data collected on the basis of machine learning, stores the received data, and separates the type or water of the oil film through SVM (Support Vector Machine) or clustering (clustering) classification method. According to the type of the oil film, a database storing data and the data collected in real time may be matched with each other to determine components of the oil film when the oil film is generated.

The apparatus may further include an output unit configured to alarm or display a result analyzed by the oil film detection unit to a user terminal using wireless communication.

The output unit may be configured to send an alarm or display the results analyzed by the oil film detection unit to the control room, and to automatically install an adsorption cloth or strainer that can remove the oil film in real time according to a predetermined control according to the analysis result. .

According to another embodiment, a UWB-based oil film detecting method includes: transmitting a signal toward a detection target surface by using an ultra wide band (UWB) and receiving data by receiving reflected waves of the transmitted signal; And detecting the oil film on the detection target surface by analyzing the collected data, and using the UWB in a non-contact manner to transmit a signal toward the surface, which is the detection target surface, to reflect the radar reflection by water. Oil film can be detected.

The step of transmitting the signal toward the detection target surface by using the UWB and receiving the reflected wave of the transmitted signal to collect data includes the presence or absence of the oil film and the type of the oil film by using the UWB based on IR (Impulse Radio) , Information about the thickness of the oil film and the distance from the surface of the water to be detected may be acquired.

The detecting of the oil film on the detection target surface by analyzing the collected data may include processing the collected data on a machine learning basis, storing the received data, and then supporting a support vector machine (CVM) or clustering. Separating each of the types of oil film or water through a classification method; Matching the data collected in real time with a database storing data according to the type of the oil film; And determining components of the oil film when the oil film is generated through the matching.

The method may further include alarming or displaying the analysis result of detecting the oil film on the detection target surface to the user terminal using wireless communication.

According to embodiments, a UWB-based oil film detection apparatus and method capable of acquiring not only the oil film but also the distance to the water surface and the thickness of the oil film using IR-based UWB can be provided.

According to embodiments, a UWB-based oil film detecting apparatus and method for detecting oil film on a deep learning basis capable of classifying distance and types of oil film in order to maintain the distance between the surface and the equipment and to secure the oil film detection accuracy in a non-contact manner can be provided. have.

1 is a diagram schematically illustrating a UWB-based oil film detection apparatus according to an embodiment.
2 is a diagram illustrating a configuration of a UWB-based oil film detecting apparatus according to an embodiment.
3 is a flowchart illustrating a UWB-based oil film detecting method according to an exemplary embodiment.
4 is a graph illustrating a received signal of a UWB radar according to an embodiment.
5 is a diagram illustrating a connection structure of a tomography convolutional neural network (CNN) based on an input image according to an embodiment.
FIG. 6 is a diagram schematically illustrating a deep learning based oil film detector model using a UWB radar, according to an exemplary embodiment.
7 is a view schematically illustrating an oil film detection experiment environment according to an embodiment.

Hereinafter, exemplary embodiments will be described with reference to the accompanying drawings. However, the described embodiments may be modified in many different forms, and the scope of the present invention is not limited to the embodiments described below. In addition, various embodiments are provided to more fully describe the present invention to those skilled in the art. Shape and size of the elements in the drawings may be exaggerated for more clear description.

Due to the spillage of oil, direct and indirect effects take a long time to recover not only ecosystems, but also economically and socially, and cause serial damage. In order to detect oil film generated by oil spill in water quickly and accurately, much research is being done on oil film detector. There are two types of oil detection methods, contact and non-contact, and contact types include conductivity, capacitance, and microwave, and non-contact infrared, ultraviolet, and laser. ), Optical, radar, and the like. The contact type requires floating equipment on the water surface, causing problems in hygiene and maintenance, and the non-contact type needs to select an oil film detector suitable for the environment because the accuracy decreases as the distance between the surface and the measuring device increases.

Therefore, in the present embodiment, oil film detection using a deep learning-based ultra wide band (UWB) radar capable of classifying distances and types of oil film in order to maintain the distance between the surface (detection surface) and the equipment and to secure the oil film detection accuracy in a non-contact manner. Provide technology. The following embodiments provide a UWB type oil film detection technology capable of acquiring not only the determination of the oil film but also the distance to the water surface and the thickness of the oil film using an IR radio based UWB.

1 is a diagram schematically illustrating a UWB-based oil film detection apparatus according to an embodiment.

Referring to FIG. 1, the UWB-based oil film detecting apparatus 100 according to an embodiment may include a radar transceiver 110, an oil film detecting unit 120, and an output unit 130. The UWB-based oil film detecting apparatus 100 may detect and determine a signal that is changed when the oil film is generated by using the radar reflection characteristic of water of the UWB.

The radar transceiver 110 is a UWB radar that collects data. The radar transceiver 110 may collect data by transmitting a signal toward a detection target surface using a UWB and receiving a reflected wave of the transmitted signal. For example, the detection target surface may be sleep.

The UWB radar may generate a continuous wave or an impulse radio to transmit a corresponding signal toward the water surface and receive the reflected wave. At this time, the antenna may use a directional antenna or a non-directional antenna as necessary for adjusting the measurement accuracy or height of the local area. According to the radar signal transmission and reception technique method, the antenna may be a single antenna method (when using frequency modulation continuous wave) and a dual antenna method in which the transmission and reception antennas are separated.

It is adjustable for optimal reception of the UWB signal via a 3-axis adjustable mobile device for changing the height or direction of the water surface. Looking at this in detail, in the case of automatic position adjustment by changing the height of the surface, the optimal position can be estimated by measuring the distance between the sleep and UWB through transmission and reception. This is a technique for responding to a change in water level for clear analysis of the signal, and the angle and height can be adjusted.

The oil film detection unit 120 may analyze the collected data to detect the oil film and determine the type of the oil film.

The received data can be classified in the following way.

All signals acquired from the receiver can be compared with a separately prepared database (Database, DB). First, the signal stored in the database represents the received signals collected in advance according to each oil type (type of oil or type of oil film). In order to classify all received signals according to oil types and classify them, oil and water may be separated through classification methods such as support vector machine (SVM) and clustering. Through this, the produced database 140 and the received signals collected in real time may be matched with each other to determine components of the oil film when water or oil film is generated.

The output unit 130 may inform the user of the analysis result through a notification and / or a display.

The determined oil film information is transmitted to the control room or managers who control the oil film, and an adsorption cloth, a sieve, etc., which can remove the oil film in real time, may be automatically installed. This can be applied to drainage facilities such as ships, factories, etc. where oil can be leaked in a habitual manner. This is applied for quick response after the discovery of oil film, and to minimize contamination before manpower response.

The alarm is based on wireless transmission based on the IoT using LTE, 5G network, etc., and can also be transmitted by wire due to short-range or communication environment limitations. In the wireless case, an alarm may be transmitted to a portable device such as a control room or a manager's smartphone using a communication network.

2 is a diagram illustrating a configuration of a UWB-based oil film detecting apparatus according to an embodiment.

Referring to FIG. 2, the UWB-based oil film detecting apparatus 100 according to an embodiment may include a radar transceiver 110, an oil film detecting unit 120, and an output unit 130 as described in FIG. 1. .

The radar transceiver 110 may collect data by transmitting a signal toward a detection target surface using UWB and receiving a reflected wave of the transmitted signal. Here, the transmitter 111 of the radar transceiver 110 may generate a UWB continuous wave or an impulse and transmit it toward the detection target surface, and the receiver 112 may receive the reflected wave after transmission, and at this time, the receiver 112 may measure signal strength over time.

The radar transceiver 110 may use a single antenna method capable of transmitting and receiving signals at the same time. In addition, the radar transceiver 110 may be a dual antenna method in which the transmitting and receiving antennas are separated.

The radar transceiver 110 may detect the oil film using the radar reflection characteristic of water as the signal is transmitted toward the surface of the detection target surface using the UWB in a non-contact manner. In particular, the radar transceiver 110 may acquire the presence or absence of an oil film, the type of the oil film, the thickness of the oil film, and the distance from the water surface as the detection target surface using the IR-based UWB.

The radar transceiver 110 may use a directional antenna or an omnidirectional antenna as necessary to adjust the measurement accuracy or height of the localized portion. In addition, the radar transceiver 110 is configured in a mobile device capable of three-axis adjustment for changing the height or direction of the water surface to be detected, the automatic position adjustment according to the change in the height of the water for optimal transmission and reception of the UWB You can make this possible.

The oil film detector 120 may detect the oil film on the detection target surface by analyzing data collected by the radar transceiver 110.

The oil film detector 120 maintains the distance between the radar transceiver 110 and the surface of the surface to be detected and secures the detection accuracy of the oil film in a non-contact manner. Can be detected.

For example, the oil film detector 120 uses the UWB radar film detection signal collected based on a deep learning-based oil film detector model using a UWB radar composed of a 1D convolutional neural network (CNN) among deep learning methods. By analyzing, the kind and thickness of an oil film can be discriminated. The deep learning based oil film detector model using convolutional neural network (CNN) and UWB radar will be described in more detail below.

The oil film detector 120 may process the collected data 121 based on the machine learning, and may store the received data and then separate the type or water of the oil film through SVM or clustering classification. The oil film detector 120 may match the database 140 in which data is stored with the data collected in real time with each other according to the type of the oil film, thereby determining components 123 of the oil film when the oil film is generated.

The output unit 130 may alarm or display the result analyzed by the oil film detection unit 120 to the user terminal using wireless communication.

The output unit 130 may transmit the result of the analysis through the differential alarm 131 according to wired / wireless and priority to the user terminal. The output unit 130 may be linked with the display device of the control room (central control room) to display the analysis result. Can be. For example, the output unit 130 transmits an alarm or a display of the result analyzed by the oil film detector 120 to the control room, and displays the adsorption cloth or strainer that can remove the oil film in real time according to a predetermined control according to the analysis result. You can have it installed automatically.

3 is a flowchart illustrating a UWB-based oil film detecting method according to an exemplary embodiment.

Referring to FIG. 3, in the UWB-based oil film detecting method according to an exemplary embodiment, a method of collecting a data by transmitting a signal toward a detection target surface using UWB, receiving a reflected wave of the transmitted signal, and collecting data And analyzing the data to detect the oil film on the detection target surface (320).

The method may further include alarming or displaying the analysis result of detecting the oil film on the detection target surface to the user terminal using wireless communication (330).

According to the embodiment, the UWB is used to transmit a signal toward the surface of the surface to be detected by using a non-contact method, and thus the oil film may be detected by using the radar reflection characteristic of water.

Hereinafter, a UWB-based oil film detection method according to an embodiment will be described in more detail with an example.

The UWB-based oil film detection method according to an embodiment may be described in more detail by using the UWB-based oil film detection apparatus according to the embodiments described with reference to FIGS. 1 and 2. Here, the UWB-based oil film detection apparatus according to an embodiment may include a transceiver 110, an oil film detector 120 and an output unit 130.

In operation 310, the radar transceiver 110 may collect data by transmitting a signal toward a detection target surface using UWB and receiving a reflected wave of the transmitted signal.

Here, the radar transceiver 110 may acquire the presence or absence of the oil film, the type of the oil film, the thickness of the oil film, and the distance to the water surface as the detection target surface by using the IR-based UWB in a non-contact manner.

In operation 320, the oil film detector 120 may detect the oil film on the detection target surface by analyzing the collected data.

The oil film detection unit 120 may detect the oil film based on a deep learning method capable of classifying distance information and types of the oil film in order to maintain the distance between the radar transceiver and the surface of the surface to be detected and to ensure the detection accuracy of the oil film in a non-contact manner. Can be. Here, deep learning may analyze data collected through a deep learning based oil film detector model using a UWB radar composed of a one-dimensional convolutional neural network (CNN).

More specifically, the oil film detection unit 120 may process the data collected on the basis of machine learning, store the received data, and then separate the type or water of the oil film through SVM or clustering classification. Thereafter, the oil film detector 120 may match the database in which data is stored and the data collected in real time according to the type of the oil film. Next, the oil film detector 120 may determine components of the oil film when the oil film is generated through matching. Can be.

In operation 330, the outputter 130 may alarm or display the analysis result of detecting the oil film on the detection target surface to the user terminal using wireless communication.

The output unit may alarm or display the result analyzed by the oil film detection unit to the control room, and automatically install an adsorption cloth or sieve that can remove the oil film in real time according to a predetermined control according to the analysis result.

Hereinafter, the UWB radar and deep learning-based oil film detection techniques used in the UWB-based oil film detection apparatus and method according to an embodiment will be described in more detail.

UWB is a wireless communication technology that uses a frequency band of more than 500MHz, which is a much wider band than the spectrum, and transmits a large amount of information at low power. In the case of IR-UWB (Impulse Radio-Ultra Wide Band), the pulse is very short, so the distance can be measured with a few centimeters of precision, so it is widely used for positioning, personnel count, non-contact breathing and heart rate measurement, and lifesaving in disaster .

P410 RCM equipment is used here, and the P410 module is a wireless transceiver with an operating frequency of 3.1 GHz to 5.3 GHz. The repetition rate for the transmitted pulse is 10 MHz and the received data is sampled at 61 ps. While the receiver is connected to the host PC for data capture, the transmitter is configured to generate short pulses continuously. The P410 was used as a monostatic radar, and the antenna used in the experiment was omni-directional, with a linear phase response, a voltage standing wave ratio of 1.75: 1 and a gain of 3dBi. UWB Radar can detect and measure distance using Time of Arrival (TOA) method and calculate range based on the received straight path signal.

4 is a graph illustrating a received signal of a UWB radar according to an embodiment.

Referring to FIG. 4, a signal received by radiating a pulse of a UWB radar is illustrated, and the x axis represents a delay time and the y axis represents an intensity ratio of the received signal. In the waveform of the actual signal, a multipath signal exists behind the linear path signal, and a weak signal or a multipath signal of various intensities is received compared to the linear path signal, which may cause an error when measuring a target distance. The mathematical model of can be expressed as

[Equation 1]

와

는 수신 신호 내에서 i 번째 샘플의 진폭과 전달 지연을 나타내며, p(t)는 신호 파형, n(t)는 주변으로부터 수신된 잡음을 의미한다. Where m ( t ) is the signal received reflected by the impulse,

Wow

Is the amplitude and propagation delay of the i th sample in the received signal, p ( t ) is the signal waveform, and n ( t ) is the noise received from the surroundings.

Since these errors vary with the medium at the surface, the characteristics will vary with each oil type.

Recently, deep learning, which shows high performance in the pattern recognition field, is a method of stacking artificial neural networks using a model composed of multiple layers of a fully connected network (FCN) layer. However, the model should be designed by composing the layers according to the characteristics of the input signal pattern. The composed layer can perform various operations through weight learning, but it is vulnerable to the recognition of local characteristics because the computation amount is very high and the global characteristics are learned. To solve this problem, we use a convolutional neural network (CNN), and it performs recursive multiplication and subsampling to recognize local and global features. The convolutional neural network (CNN) is suitable because the nature of the UWB signal requires recognition of the local part.

5 is a diagram illustrating a connection structure of a tomography convolutional neural network (CNN) based on an input image according to an embodiment.

Referring to FIG. 5, x _i 510 is input information, y _i 530 is output information, w ₁ , w _{2 in} a connection structure 500 of a tomographic convolutional neural network (CNN) based on an input image. And w ₃ 520 means the weight of each line. The result of FIG. 5 can be expressed as follows.

[Equation 2]

Since the image is two-dimensional, it can be calculated by adding an axis to Equation 2. As shown in FIG. 5 and Equation 2, the convolutional neural network (CNN) learns based only on the surrounding pixel information. In addition, the convolutional neural network (CNN) reduces the amount of computation by applying the same weight in w ₁ , w _2, and w ₃ in other areas, and simultaneously applies the local characteristics of one region to distant regions.

However, since a single weight map can extract only one feature, a weighting group that functions as a filter is needed to classify various features. The weight map corresponding to one filter is called a kernel, and the number of kernels is equal to the number of filters. Therefore, the convolutional neural network (CNN) can learn by configuring various kernels corresponding to each extracted feature to extract various features.

FIG. 6 is a diagram schematically illustrating a deep learning based oil film detector model using a UWB radar, according to an exemplary embodiment.

Referring to FIG. 6, a UWB-based oil film detecting apparatus according to an embodiment detects UWB radar film collected based on a deep learning-based oil film detector model 600 using a UWB radar configured with a 1-D convolutional neural network (CNN). You can analyze the signal.

Convolutional neural networks (CNNs), on the other hand, generally have three types of layers. The convolution layer 620 is a layer for extracting convolutional features. The convolution layer 620 is a layer for extracting meaningful features. The pooling layer is characterized by a large number of pixels due to the characteristics of the image. The subsampling layer is used to reduce the feedforward layer, and the feedforward layer is a layer classified using the features derived from the convolutional layer and the pooling layer.

The deep learning based oil film detector model 600 using the UWB radar not only inputs the general time domain UWB signal 611 to the first channel but also receives the preprocessed signal 612 through the FFT for extracting the characteristics of the frequency domain. Input into the second channel allows simultaneous analysis of the time domain and frequency domain. The output of the deep learning based oil film detector model 600 using the UWB radar follows a one-hot encoded data structure.

The deep learning based oil film detector model 600 using the UWB radar may include five convolution layers 620 and a pulling layer installed at each end of the layer. In addition, a neural network including a full connect 630 may be configured to classify types and distances of an oil film at the output of the deep learning-based oil film detector model 600 using the UWB radar.

Accordingly, the type and thickness of the oil film may be determined (640).

7 is a view schematically illustrating an oil film detection experiment environment according to an embodiment.

Referring to FIG. 7, the distance of the UWB module 710 which transmits and receives water and UWB 711 in the experimental environment to measure the distance and the type of oil film is measured at 10 cm intervals from 0.5 m to 1.5 m and a tank ( The horizontal, vertical and height of the 720 is 0.3m and the height of the water 721 is 0.15m so that a signal separated from the reflection of the water tank 720 may be spaced apart from the height of the bottom of the water tank 720. In addition, the thickness of the oil film 722 is set to 0.006 m.

Here, the oil to be detected to detect the oil film 722 is analyzed using kerosene, diesel oil, edible oil and water to analyze the detection accuracy. In order to verify the accuracy and validity of the proposed UWB-based oil film detection device, it is verified by Cross Validation.Training dataset is 80% and Test and Validation is 20%. Set to%.

Experimental results show that the proposed UWB-based oil film detection device has high accuracy and distance, and proves the possibility of detection technique.

The apparatus described above may be implemented as a hardware component, a software component, and / or a combination of hardware components and software components. For example, the devices and components described in the embodiments include, for example, processors, controllers, arithmetic logic units (ALUs), digital signal processors, microcomputers, field programmable arrays (FPAs), It may be implemented using one or more general purpose or special purpose computers, such as a programmable logic unit (PLU), microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to the execution of the software. For convenience of explanation, one processing device may be described as being used, but one of ordinary skill in the art will appreciate that the processing device includes a plurality of processing elements and / or a plurality of types of processing elements. It can be seen that it may include. For example, the processing device may include a plurality of processors or one processor and one controller. In addition, other processing configurations are possible, such as parallel processors.

The software may include a computer program, code, instructions, or a combination of one or more of the above, and configure the processing device to operate as desired, or process it independently or collectively. You can command the device. Software and / or data may be any type of machine, component, physical device, virtual equipment, computer storage medium or device in order to be interpreted by or to provide instructions or data to the processing device. It can be embodied in. The software may be distributed over networked computer systems so that they may be stored or executed in a distributed manner. Software and data may be stored on one or more computer readable recording media.

The method according to the embodiment may be embodied in the form of program instructions that can be executed by various computer means and recorded in a computer readable medium. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the media may be those specially designed and constructed for the purposes of the embodiments, or they may be of the kind well-known and available to those having skill in the computer software arts. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tape, optical media such as CD-ROMs, DVDs, and magnetic disks, such as floppy disks. Magneto-optical media, and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine code generated by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like.

Although the embodiments have been described by the limited embodiments and the drawings as described above, various modifications and variations are possible to those skilled in the art from the above description. For example, the described techniques may be performed in a different order than the described method, and / or components of the described systems, structures, devices, circuits, etc. may be combined or combined in a different form than the described method, or other components. Or even if replaced or substituted by equivalents, an appropriate result can be achieved.

Therefore, other implementations, other embodiments, and equivalents to the claims are within the scope of the claims that follow.

Claims (6)

A radar transceiver configured to transmit a signal toward a detection target surface by using an ultra wide band (UWB) and receive reflected waves of the transmitted signal to collect data; And
Oil film detection unit for detecting the oil film on the detection target surface by analyzing the data collected by the radar transceiver
Including,
The radar transceiver,
Using the UWB based on the impulse radio (IR) in a non-contact manner, as the signal is transmitted toward the surface of the detection target surface, the oil film is detected using the radar reflection characteristic of water.
The oil film detection unit,
In order to maintain the distance between the radar transceiver and the water surface being the detection target surface in a non-contact manner and to secure the detection accuracy of the oil film, the oil film is based on deep learning based on distance learning and the type of oil film. By using the UWB, the presence or absence of the oil film, the type of the oil film, the thickness of the oil film, and distance information with respect to the water surface being the detection target surface are acquired by using the UWB.
The deep learning analyzes the data collected through a deep learning based oil film detector model using a UWB radar composed of a 1D convolutional neural network (CNN), and uses a deep learning based film detector using the UWB radar. The model inputs the UWB signal in the time domain to the first channel, analyzes the time domain and the frequency domain simultaneously by inputting the preprocessed signal through the FFT to the second channel to extract the features of the frequency domain, and the one-hot method. Output according to the encoded data structure
UWB-based oil film detection device, characterized in that. delete The method of claim 1,
The radar transceiver,
It uses directional antenna or omnidirectional antenna to adjust the measurement accuracy or height of the local area.It uses single antenna method that can transmit and receive signals at the same time, and 3-axis adjustment is used to change the height or direction of the water surface to be detected. Configured in a mobile device capable of automatic positioning according to the height change of the water surface for optimal transmission and reception of the UWB
UWB-based oil film detection device, characterized in that. delete The method of claim 1,
The oil film detection unit,
After processing the data collected on the basis of machine learning, and storing the received data, the type of the oil film or the water is separated through the SVM (Support Vector Machine) or clustering (clustering) classification method, respectively, the type of the oil film Matching the data collected in real time with the database in which data is stored, and determining components of the oil film when the oil film is generated.
UWB-based oil film detection device, characterized in that. The method of claim 1,
Output unit for alarm or display the result analyzed by the oil film detection unit to the user terminal using wireless communication
More,
The output unit,
The result of the analysis by the oil film detection unit is transmitted to the control room for alarm or display, and the adsorption cloth or sieve which can remove the oil film in real time according to the preset control according to the analysis result is automatically installed.
UWB-based oil film detection device, characterized in that.

Images