KR101558438B1 - Method and Apparatus for Active Target Classification and Feature Vector Extraction Using Fractional Fourier Transform - Google Patents

Abstract

The present invention relates to a target object recognizing technology, more specifically, to a method for receiving a scattered signal through a distributed sensor by using an active sonar signal or a radar signal in an active target identifying field to recognize a target object and a device thereof. The method comprises: a signal inputting step of receiving a receiving signal from a target; a detecting step of detecting a partial Fourier transfer area existing in a target signal; an extracting step of extracting a feature vector from the detected partial Fourier transfer area; and a target recognizing step of recognizing the target about the feature vector extracted by using a recognizer.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to a method and an apparatus for extracting a feature vector using a partial Fourier transform and a method and apparatus for identifying an active target using an Fourier transform,

The present invention relates to a target object recognition technique, and more particularly, to a method and device for recognizing a target object by receiving a scattered signal using an active sonar signal or a radar signal among active target identification fields through a dispersion sensor, Lt; / RTI >

Particularly, the present invention performs partial Fourier transform on an amplified received signal, detects a meaningful signal region in a partial Fourier transform domain, extracts a feature vector from the detected domain, and uses the extracted feature vector as an input to the recognizer To a method and apparatus for recognizing a target.

Since active sonar and target recognition use only reflected echoes from the target, it is impossible to analyze the noise source of the target, and it is necessary to first identify the target by analyzing the pattern of the reverberation according to the shape and size, Research is needed in the direction of more accurately confirming the class of the target by using a Doppler and a maneuvering speed.

Characteristic factors that can be extracted from the active target signal in a multistatic sonar system include Doppler and maneuver speed. The target activation speed is determined by the Doppler and attitude angle of the target, and the target Doppler can be estimated by the continuous wave (CW) pulse of the active sonar.

A basic study for the recognition of active targets is to use the correlation of the beam segmentation which can estimate the target length and attitude angle, the technique to improve the detection and recognition performance of the active sonar in the reverberant reverberation environment and the DORT (Time Reversal Operator Decomposition (HMM) and Hidden Markov models (HMM).

1. Korean Registered Patent No. 10-1103493 2. Japanese Patent No. 4977042 3. U.S. Patent No. 7180442

1. "LFM Signal Separation Using Fractional Fourier Transform", International Journal of The Institute of Information and Telecommunication Volume 17, No. 3, March 2013, pp.540-545.

The present invention is conceived to solve the need according to the above background art, and it is an object of the present invention to provide a method and apparatus for receiving a signal scattered by an active sonar signal or a radar signal among active target identification fields through a dispersion sensor, And a method and apparatus for vector extraction / active target identification.

In order to achieve the above object, the present invention provides a method for extracting a feature vector / active target identification method for recognizing a target object by receiving a signal scattered by an active sonar signal or a radar signal among active target identification fields through a dispersion sensor, .

The identification method comprises:

A signal input step of inputting a reception signal from a target;

A detection step of detecting a partial Fourier transform domain in which a target signal exists using a partial Fourier transform on a received signal inputted at a predetermined angle or at a predetermined distance;

An extraction step of extracting a feature vector from the detected partial Fourier transform domain; And

And a target recognition step of recognizing the target with respect to the extracted feature vector.

In this case, the received signal may be a plurality of signals continuously spread over a predetermined angle or a predetermined distance from the target, or a signal spread from the target.

The detecting step may include: setting a reference threshold Th and a magnitude sustain period Td; Determining as a start portion of a signal to be detected if a signal component that is higher than the reference threshold value and is greater than or equal to the magnitude sustain period continues; And terminating the detection if the magnitude is again smaller than the reference threshold value and the signal component of the magnitude sustain period (Td) or more does not exist after the start of the signal is started.

The extracting step may include: obtaining a magnitude in a partial Fourier transform domain; Dividing a partial Fourier transform domain into a plurality of (M) subbands; And extracting a feature vector for each subband.

The target recognition step may include recognizing a target using a feature vector extracted from the partial Fourier transform domain as an input.

The reception signal may be received using a dispersion sensor.

The received signal may be a signal spread from a target after radiating a continuous wave (CW) signal or a linear frequency modulated wave (LFM) signal with a different signal length depending on an object or distance to be detected have.

In addition, a method of extracting the feature vector includes a method of extracting an average energy component of each interval by dividing the partial Fourier transform domain into subbands, which are small intervals, and a method of extracting only a value of a subband having a certain threshold value or more, And a method of selecting a local maximum component in a partial Fourier transform domain as a feature vector.

In addition, the recognizer may be an artificial intelligence network recognizer or a SVM (Support Vector Machine) recognizer.

On the other hand, another embodiment of the present invention includes an input unit for inputting a received signal from a target; A partial Fourier transform unit applying a partial Fourier transform to a received signal inputted at a predetermined angle or at a predetermined distance; A signal detector for detecting a partial Fourier transform domain in which a target signal exists; A feature extraction unit for extracting a feature vector from the detected partial Fourier transform domain; And a recognizer that recognizes the target with respect to the extracted feature vector. The feature vector extracting and the active target identifying device using the partial Fourier transform are provided.

According to the present invention, only a meaningful signal component is extracted in a partial Fourier transform domain using a partial Fourier transform in feature vector extraction and active target identification using a partial Fourier transform, the detected signal domain is divided into a plurality of subbands, By extracting the features in various ways and using them as input to the recognizer, it is possible to reflect the characteristics of the received target signal according to the shape / material of the target.

Another advantage of the present invention is that the target can be more efficiently identified because it reflects the characteristics of the received target signal according to the shape / material of the target.

1 is an extraction identification device 100 for performing feature vector extraction and active target identification using partial Fourier transform according to an embodiment of the present invention.
FIG. 2 is a detailed block diagram of the feature extraction unit 124 for extracting the feature vector shown in FIG.
FIG. 3 is a flowchart illustrating a process of performing feature vector extraction and active target identification using partial Fourier transform according to an exemplary embodiment of the present invention. Referring to FIG.
FIG. 4 is a flowchart illustrating the feature vector extraction step (S330) shown in FIG. 3 in detail.
5 and 6 are conceptual diagrams illustrating an example of a LFM signal used for target recognition according to an embodiment of the present invention in a time domain and a frequency domain.
7 is a conceptual diagram for explaining a general basic Fourier transform characteristic.
8 is a conceptual diagram showing the shape of two LFM signals in a general Fourier transform domain.
FIG. 9 is a graph showing an example in which only the signal region is detected by the signal detecting unit 123 shown in FIG.
FIG. 10 is a diagram showing the detected signal region 1010 shown in FIG.
11 and 12 are conceptual diagrams illustrating the concept of dividing a signal region into subbands of a small unit in the feature extraction unit 124 shown in FIG.
FIG. 13 is a conceptual diagram illustrating an example of extracting local maxima during feature extraction for M subbands in the calculation unit 230 shown in FIG.

While the invention is susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and will herein be described in detail. It is to be understood, however, that the invention is not to be limited to the specific embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention.

Like reference numerals are used for similar elements in describing each drawing.

The terms first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another.

For example, without departing from the scope of the present invention, the first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component. The term "and / or" includes any combination of a plurality of related listed items or any of a plurality of related listed items.

Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

Terms such as those defined in commonly used dictionaries are to be interpreted as having a meaning consistent with the contextual meaning of the related art and are to be interpreted as either ideal or overly formal in the sense of the present application Should not.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a method and apparatus for extracting a feature vector and an active target identifying method using a partial Fourier transform according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

1 is an extraction identification device 100 for performing feature vector extraction and active target identification using partial Fourier transform according to an embodiment of the present invention. Referring to FIG. 1, the extraction identification apparatus 100 includes an input unit 110 for receiving a received signal, a partial Fourier transform applied to the received signal to detect an area in which a scattered signal exists, And a recognizer 130 for recognizing a target on the extracted feature vector and calculating a recognition result.

Here, the signal processing unit 120 includes a partial Fourier transform unit 121 for applying a partial Fourier transform to a received signal inputted at a certain angle or a predetermined distance, a signal detecting unit 123 for detecting a partial Fourier transform region in which a target signal exists And a feature extraction unit 124 for extracting a feature vector from the detected partial Fourier transform domain.

Although the signal processor 120 and the recognizer 130 are shown separately in FIG. 1, the recognizer 130 may be incorporated in the signal processor 120. FIG.

The input unit 110 receives a received signal that is continuously spread over a predetermined angle or a predetermined distance. The received signal becomes an active sonar signal. The active sonar signal radiates CW (Continuous Wave) or LFM (Linear Frequency Modulated wave) signals with different signal lengths depending on the object and / or distance to be detected, And recognizes the target based on the spread signal.

만큼 회전시킨 결과를 의미한다.The partial Fourier transform unit 121 performs partial Fourier transform on the input signal. The partial Fourier transform is a generalized form Fourier transform. This partial Fourier transform has different results depending on the angle with the time axis in the time-frequency domain. In other words, a general Fourier transform is performed in the time-frequency domain from the time axis to the frequency axis

As shown in FIG.

는 다음식에 의해 구할 수 있다. On the other hand, the partial Fourier transform can obtain different analysis results depending on the angle with the time axis. When an LFM (Linear Frequency Modulated wave) signal is targeted, an optimal conversion parameter can be obtained according to the chirp rate of the LFM signal. Optimal conversion parameter

Can be obtained by the following.

는 LFM 신호의 첩 속도(chirp rate)와 관련된 파라미터이고,

는 수신된 신호의 표본화 주파수이며,

은 수신된 신호의 샘플 수를 각각 나타낸다.here,

Is a parameter related to the chirp rate of the LFM signal,

Is the sampling frequency of the received signal,

Represents the number of samples of the received signal.

The signal detecting unit 123 detects only a region in which a significant signal (a signal region judged to exist a target signal) exists in the partial Fourier transform domain. The reference threshold value Th and the magnitude sustain period Td are set and set as the start of the signal to be detected if the signal component higher than the reference threshold value and longer than the magnitude sustain period continues.

After the beginning of the signal, the detection is terminated if the magnitude is again less than the reference threshold and there is no signal component greater than or equal to the magnitude sustained interval Td.

FIGS. 9 and 10 show the detected signal components after receiving the scattered signals in the target and performing the starting point detection by applying the partial Fourier transform.

1, the recognizer 130 actually performs target recognition using the feature vector extracted by the feature extraction unit 124 as an input. As the recognizer 130, a recognizer such as an artificial intelligence network (Neural Network) or SVM (Support Vector Machine) is used. However, the present invention is not limited to this specific recognition algorithm, and various recognition methods are possible. In this case, the artificial intelligence recognizer is an identifier for recognizing an object to be recognized according to a manner in which a human neural network model human neurons. The SVM recognizer recognizes a recognition object, It is an identifier that learns and recognizes by finding the most distant plane.

FIG. 2 is a detailed block diagram of the feature extraction unit 124 for extracting the feature vector shown in FIG. 2, the feature extraction unit 124 receives a region corresponding to a signal component detected by the signal detection unit 123 (see FIG. 1), calculates a magnitude in the partial Fourier transform region, A subband dividing unit 220 for dividing the input signal into M subbands, a calculating unit 230 for extracting feature vectors for each subband, and the like.

In the subband division unit 220, the entire signal region is divided into M subbands. At this time, a method of dividing into subbands is performed by using a triangular filter or a rectangular filter. FIG. 11 is a conceptual diagram of dividing an entire signal region using a triangular filter, and FIG. 12 is a view showing a state where an entire signal region is divided using a rectangular filter.

The calculating unit 230 can extract feature vectors in three ways. In the first method, the partial Fourier transform domain is divided into subbands, which are small intervals, and the average energy component of each interval is extracted, and the energy of each subband is used as the input of the recognizer as the feature vector.

In the second scheme, only a value in a subband that is equal to or greater than a predetermined threshold value is selected as a feature vector and used as an input to the recognizer 130.

Finally, the third scheme is to select a local maximum component in the partial Fourier transform domain as a feature vector and use it as an input to the recognizer 130. 13 shows the local maximum component in the partial Fourier transform domain. This will be described later.

FIG. 3 is a flowchart illustrating a process of performing feature vector extraction and active target identification using partial Fourier transform according to an exemplary embodiment of the present invention. Referring to FIG. Referring to FIG. 3, a signal dispersed from a target is received by using a dispersion sensor (not shown) and is used as an input signal (step S310). In this case, the received signals scattered on the target are not only subjected to only one signal spread in the target, but also transmitted continuously at a constant angle and / or at a constant distance using a plurality of dispersion sensors arranged at a certain angle and / As the input signal.

A partial Fourier transform is performed on an input signal spread over a certain angle or a predetermined distance to detect a partial Fourier transform region in which a signal exists (step S320). The detection of the partial Fourier transform domain is a process of detecting a region where a signal exists with respect to an input signal. The reference threshold value Th and the magnitude sustain period Td are set so that a signal component higher than the reference threshold value, If it is sustained, it is set as the beginning of the signal to be detected. Also, after the beginning of the signal, the detection is terminated if the magnitude is again less than the reference threshold and there is no signal component greater than or equal to the magnitude sustained interval Td.

Then, a feature vector is extracted from the detected partial Fourier transform domain (step S330).

When the feature vector is extracted, the extracted feature vector is input to the recognizer (130 in FIG. 1) to recognize the target (step S340).

FIG. 4 is a flowchart illustrating the feature vector extraction step (S330) shown in FIG. 3 in detail. Referring to FIG. 4, a magnitude in the partial Fourier transform domain is obtained (step S410).

It is further divided into M subbands (step S420).

After the subband division, a feature vector is extracted for each subband (step S430).

5 and 6 are conceptual diagrams illustrating an example of a LFM signal used for target recognition according to an embodiment of the present invention in a time domain and a frequency domain. FIG. 5 shows the amplitude of the signal in the time domain, and FIG. 6 shows the magnitude of the signal 610 in the frequency domain.

7 is a conceptual diagram for explaining a general basic Fourier transform characteristic. Referring to FIG. 7, the shape of two LFM signals in a general Fourier transform domain is shown.

FIG. 8 is a conceptual diagram showing the shape of two LFM (Linear Frequency Modulated Wave) signals in a partial Fourier transform domain. Referring to FIG. 8, the shape of two LFM signals in the partial Fourier transform domain is shown. Further, an amplitude relation according to the frequency is shown in the case of a = 0.95 (810), a = 0.952 (820), a = 0.954 (830), and a = 0.957 (840).

FIG. 9 is a graph showing an example in which only the signal region is detected by the signal detecting unit 123 shown in FIG. FIG. 10 is a diagram showing the detected signal region 1010 shown in FIG. Particularly, FIGS. 9 and 10 show the detected signal components after receiving the scattered signal in the target and performing the start point detection by applying the partial Fourier transform.

11 and 12 are conceptual diagrams illustrating the concept of dividing a signal region into subbands of a small unit in the feature extraction unit 124 shown in FIG. Particularly, FIG. 11 is a conceptual diagram for dividing an entire signal region using a triangular filter 1110, and FIG. 12 shows a shape for dividing an entire signal region using a rectangular filter 1210 Giving.

FIG. 13 is a conceptual diagram illustrating an example of extracting local maxima during feature extraction for M subbands in the calculation unit 230 shown in FIG. FIG. 13 shows a local maximum component 1321 in the partial Fourier transform domain 1310. FIG.

In addition, it will be appreciated that the various illustrative logical blocks, and algorithm steps described in connection with the embodiments disclosed herein, may be implemented as electronic hardware, computer software, or combinations thereof. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality.

Whether such functionality is implemented as hardware or software depends upon the design constraints and specific applications imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the exemplary embodiments of the present invention.

The various illustrative logical blocks described in connection with the embodiments disclosed herein may be implemented or performed with a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) Discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein.

A general purpose processor may be a microprocessor, but, in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. A processor may also be implemented as a combination of computing devices, e.g., a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration.

The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module may reside in a memory such as a random access memory (RAM), a flash memory, a read only memory (ROM), an electrically programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), registers, CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor such that the processor can read information from, and write information to, the storage medium. Alternatively, the storage medium may be integral to the processor. The processor and the storage medium may reside in an ASIC. The ASIC may reside in a user terminal. Alternatively, the processor and the storage medium may reside as discrete components in a user terminal.

In one or more exemplary embodiments, the functions described may be implemented in hardware, software, firmware, or any combination thereof. When implemented in software, the functions may be stored or transmitted as one or more instructions or code on a computer-readable medium. The computer-readable medium includes both a communication medium and a computer storage medium, including any medium that facilitates transfer of a computer program from one place to another. The storage medium may be any available media that can be accessed by a computer. By way of example, and not limitation, such computer-readable media can comprise any form of computer readable medium, such as RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, Or any other medium which can be used to carry or store the program code and which can be accessed by a computer.

Also, any connection is properly termed a computer-readable medium. For example, if the software is transmitted from a web site, server, or other remote source using wireless technologies such as coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or infrared, wireless and microwave, Wireless technologies such as cable, fiber optic cable, twisted pair, DSL, or infrared, radio, and microwave are included within the definition of media. As used herein, discs and discs include compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disc and Blu-ray disc, ) Typically reproduce data autonomously, but discs use lasers to optically reproduce the data. Also, the above combinations should be included within the scope of computer-readable media.

The previous description of the disclosed exemplary embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these exemplary embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Accordingly, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

100: extraction identification device
120: Signal processor
121: partial Fourier transform unit
123:
124: Feature extraction unit
130: recognizer
210:
220: minute installment
230:

Claims (10)

A signal input step of inputting a reception signal from a target;
A detecting step of detecting a partial Fourier transform domain in which a target signal exists by using a partial Fourier transform on a received signal inputted at a certain angle or a certain distance;
An extraction step of extracting a feature vector from the detected partial Fourier transform domain; And
And a target recognition step of recognizing a target on a feature vector extracted using a recognizer,
In the method of extracting the feature vector, a method of dividing a partial Fourier transform domain into subbands, which are small intervals, to extract an average energy component of each interval, and a method of selecting only a value that is equal to or greater than a certain threshold value in a subband And a method of selecting a local maximum component in a partial Fourier transform domain as a feature vector. The feature vector extracting method and the active target identification method using the partial Fourier transform.
The method according to claim 1,
Wherein the received signal is a plurality of signals that are continuously scattered at a certain angle or a predetermined distance from the target, or a signal that is spread from the target, and extracting the feature vector using the partial Fourier transform.
The method according to claim 1,
Wherein the detecting step comprises:
Setting a threshold Th and a sustain period Td;
Determining as a start portion of a signal to be detected if a signal component that is higher than the reference threshold value and is greater than or equal to the magnitude sustain period continues; And
And after the start of the signal, the detection is terminated if the magnitude is again smaller than the reference threshold and there is no signal component greater than or equal to the magnitude sustain period (Td). Extraction and active target identification methods.
The method according to claim 1,
Wherein the extracting step comprises:
Obtaining a magnitude in a partial Fourier transform domain;
Dividing a partial Fourier transform domain into a plurality of (M) subbands; And
Extracting a feature vector for each subband; and extracting a feature vector and performing an active target identification using a partial Fourier transform.
The method according to claim 1,
The target recognition step may include:
And recognizing a target using the feature vector extracted from the partial Fourier transform domain as an input. The feature vector extracting method and the active target identifying method using the partial Fourier transform.
The method according to claim 1,
Wherein the received signal is received using a distributed sensor. ≪ RTI ID = 0.0 > [10] < / RTI >
The method according to claim 1,
Wherein the received signal is a signal spread from a target after a CW (Continuous Wave) or LFM (Linear Frequency Modulated wave) signal is emitted with a different signal length depending on an object or distance to be detected, Feature Vector Extraction and Active Target Identification Using.
delete The method according to claim 1,
Wherein the recognizer is an artificial intelligence network recognizer or a support vector machine (SVM) recognizer, and extracting feature vectors and performing active target identification using partial Fourier transform.
An input unit for inputting a reception signal from the target;
A partial Fourier transform unit applying a partial Fourier transform to a received signal inputted at a predetermined angle or at a predetermined distance;
A signal detector for detecting a partial Fourier transform domain in which a target signal exists;
A feature extraction unit for extracting a feature vector from the detected partial Fourier transform domain; And
And a recognizer for recognizing the target with respect to the extracted feature vector,
In the method of extracting the feature vector, a method of dividing a partial Fourier transform domain into subbands, which are small intervals, to extract an average energy component of each interval, and a method of selecting only a value that is equal to or greater than a certain threshold value in a subband And a method of selecting a local maximum component in a partial Fourier transform domain as a feature vector. The feature vector extracting and active target identifying device using the partial Fourier transform.

Images