KR100991348B1 - Method for motion compensation of stepped-frequency radar images using PSO - Google Patents

Abstract

The present invention obtains a radar image by receiving a step frequency radar signal reflected from the target after being transmitted to a target side, and sets the entropy of the radar image as a cost function to correct a phase error due to the fluctuation of the target. In addition, the present invention provides a method for compensating for shaking a step frequency radar image using PSO by applying Particle Swarm Optimization (PSO) as a technique for minimizing the cost function.

According to the fluctuation compensation method of the disclosed step frequency radar image, a high quality radar image from which a phase error is removed can be obtained by applying a PSO method to the acquired step frequency radar image to compensate for the fluctuation caused by the movement of a target. There is an advantage to that. In addition, when the PSOI (Particle Swarm Optimization with an Island model) technique, in which a plurality of PSOs cooperate with each other and search for an optimal solution, can improve the performance of the shaking compensation more clearly, Radar images can be obtained.

Description

Method for motion compensation of stepped-frequency radar images using PSO}

The present invention relates to a fluctuation compensation method of a step frequency radar image using a PSO, and more particularly, to obtain a more clear radar image by correcting a phase error according to the movement of a target in a step frequency radar image. The present invention relates to a method for compensating for shaking of radar images.

The radar image acquired with respect to a target expresses the physical phenomenon reflected when an electromagnetic wave enters a target in 1-dimensional or 2-dimensional space. One-dimensional radar images include a range profile and a high resolution range profile. A two-dimensional radar image includes a SAR (Synthetic Aperture Radar) image and an ISAR ( Inverse Synthetic Aperture Radar images.

Acquisition of such radar images is the most important process required for performing non-cooperative target recognition (NCTR) or automatic target recognition (ATR), which is a method of identifying targets from radar images. . In other words, in order to perform NCTR or ATR, acquisition of a high resolution and high quality radar image is most important.

However, the tactically important target aircraft, tanks, armored vehicles, etc. are movable targets, so if the target makes a radar image while moving, the image is blurred. That is, when the target moves, the radar signal received from the target includes a phase error due to the movement of the target, thereby blurring the radar image. This phenomenon is because the radar image is affected by the phase error caused by the Doppler effect due to the movement of the target.

The method for removing the phase error from the radar image to obtain a high quality radar image is called motion compensation or autofocus technique. If the phase error caused by the movement of the target is not removed, the radar image is blurred. If the NCTR or ATR is performed using the blurred image, the desired performance cannot be obtained.

SUMMARY OF THE INVENTION An object of the present invention is to provide a method for compensating fluctuation of a step frequency radar image using PSO, which can obtain a high quality radar image from which phase error is eliminated by compensating the fluctuation of a step frequency radar image using a PSO technique. .

The present invention obtains a radar image by receiving a step frequency radar signal reflected from the target after being transmitted to a target side, and sets the entropy of the radar image as a cost function to correct a phase error due to the fluctuation of the target. In addition, the present invention provides a method for compensating for shaking a step frequency radar image using PSO by applying Particle Swarm Optimization (PSO) as a technique for minimizing the cost function.

Here, the fluctuation compensation method is a technique for minimizing the cost function, by applying a Particle Swarm Optimization with an Island model (PSOI) method, which is a method in which a plurality of PSOs, rather than one, cooperate to find an optimal solution. The shaking compensation can be performed.

According to the fluctuation compensation method of the step frequency radar image using the PSO of the present invention, a high quality radar image is obtained by removing the phase error as the fluctuation compensation by the movement of the target is performed by applying the PSO method to the acquired step frequency radar image. There is an advantage that can be obtained. In addition, when the PSOI method, which is a method in which a plurality of PSOs cooperate with each other and searches for an optimum solution, is applied, the performance of the shaking compensation can be further improved, and a clearer and clear radar image can be obtained.

1 is a configuration diagram of a chirp pulse and a step frequency radar, and FIG. 2 is a distance side view of the chirp pulse and step frequency radar of FIG. 1.

Prior to the detailed description of the present invention, the chirp pulse radar and the step frequency radar will be briefly described. Inverse Synthetic Aperture Radar (ISAR) is a technique for obtaining a two-dimensional image of a target, and is used as a characteristic vector of a target in an automatic target equation with a one-dimensional distance profile. There are two ways to obtain ISAR image. There are two methods using chirp pulse and stepped frequency.

Referring to (a) of FIG. 1, a chirp pulse radar is widely used for a general image forming radar, and transmits and receives a short pulse having a given bandwidth. The received signal is matched-filtered with a pre-stored replica. Since the length of the pulse in the chirp pulse radar is relatively short relative to the speed of the target, it is assumed that the target is stationary as soon as the transmitted pulse is reflected at the target. Thus, referring to FIG. 2, the distance side view obtained using the chirp pulse radar is well focused. However, these chirppulse radars often have the disadvantages of complex hardware and high cost. For reference, the horizontal axis of Figure 2 represents a range bin index, the vertical axis represents the size of the signal.

Referring to (b) of FIG. 1, the step frequency radar transmits a burst composed of a series of monotone pulses having one frequency, and then performs an Inverse Fourier Transform (IFFT) to obtain a distance side view. Step frequency radar has the advantage of simple hardware structure and high resolution. However, referring to FIG. 2, the image is severely blurred due to the movement of the target between pulses in the burst while one burst is reflected from the target. Several studies have proposed autofocusing of step-frequency ISAR images, but these techniques are performed on the assumption that the movement of the target is very small while one burst is reflected from the target.

In the present invention, an oscillation compensation algorithm for compensating for inter-pulse movement due to the movement (oscillation) of the target in forming the step frequency ISAR image as described above will be described in detail. The present invention can be applied to a range profile and an HRR profile, which are one-dimensional radar images, and to SAR and ISAR images, which are two-dimensional radar images.

Hereinafter, a method of compensating for shaking of a step frequency radar image according to an exemplary embodiment of the present invention will be described in detail with reference to FIGS. 3 to 6.

3 is a configuration diagram of a step frequency radar signal model used in the present invention, and FIG. 4 is a schematic flowchart of a fluctuation compensation method of a step frequency radar image according to an embodiment of the present invention. 5 is an explanatory diagram of a PSO technique applied to FIG. 4, and FIG. 6 is a configuration diagram of a PSOI technique that is an application of the PSO technique of FIG. 5.

First, referring to FIG. 4, a specific radar device (not shown) transmits the step frequency radar signal to a target side. Thereafter, the radar device receives the step frequency radar signal reflected from the target after being transmitted to the target side to acquire a radar image (S110).

개의 펄스(Pulse)들로 이루어진

개의 버스트(Burst)들을 포함하여 구성된다. 각 펄스의 반복주파수(PRF;Pulse Repetition Frequency)는

이다. 하나의 버스트 내에서 각 주파수 성분(

)은 수학식 1과 같다.Here, referring to the signal model of FIG. 3, the step frequency radar signal for obtaining the ISAR image is

Of pulses

It consists of three bursts (Burst). Pulse Repetition Frequency (PRF) of each pulse

to be. Each frequency component within one burst

) Is the same as Equation 1.

[Equation 1]

,

,

는 시작 주파수,

는 주파수 스텝을 나타낸다. 이 경우, 각 버스트의 총 밴드폭은

가 된다. 그리고, 각 버스트 시간 동안 레이더 시선(radar line-of-sight)은 고정된 것으로 가정한다. 이때, 표적으로부터 반사되어 레이더장치로 수신된 스텝주파수 레이더 신호는 수학식 2와 같다.here,

Is the starting frequency,

Denotes a frequency step. In this case, the total bandwidth of each burst

Becomes And, radar line-of-sight is assumed to be fixed during each burst time. In this case, the step frequency radar signal reflected from the target and received by the radar device is represented by Equation 2.

[Equation 2]

는 산란원의 개수,

는

번째 산란원의 크기이다. 그리고,

는 시간이

일 때 레이더 시선방향으로 투영된

번째 산란원의 거리로서, 수학식 3으로 표현된다. 물론,

이다.here,

Is the number of scattering circles,

Is

The size of the first scattering circle. And,

Has time

Projected in the radar line of sight

The distance of the second scattering circle is expressed by Equation 3 below. sure,

to be.

&Quot; (3) "

은

번째 버스트 시작 시에

번째 산란원의 초기거리이며,

이며,

과

은 각각 레이더 시선방향에 대한 속도와 가속도를 나타낸다.here,

silver

At the start of the first burst

Initial distance of the first spawning source,

,

and

Represents the speed and acceleration with respect to the radar line of sight, respectively.

However, when the target is a movable target of a tank or an armored vehicle, the acquired radar image is blurred due to a phase error caused by the movement of the target. That is, a phase error occurs according to the Doppler effect due to the movement of the target, which causes the radar image to be affected and the focal point is blurred, resulting in deterioration of the radar image quality. In addition, in the case of a radar using a step frequency waveform, such a phase error greatly affects the quality of the radar image. This low quality or low resolution radar image makes it difficult to perform NCTR or ATR target recognition.

성분은 ISAR 영상이 흐려지게 만드는 주요 요인이 되므로, 각 버스트에서 레이더 시선방향의 속도(

)와 가속도(

)를 적절히 추정하여 이를 제거해야 한다. Furthermore, referring to equation (2), between pulses

The component is the main factor in blurring the ISAR image, so the radar visual velocity (for each burst)

) And acceleration (

) Should be properly estimated and eliminated.

To this end, the radar device receives a reflected step frequency radar signal to obtain a radar image, in order to correct the phase error according to the fluctuation of the target, to set the entropy (Entropy) of the radar image as a cost function In addition, by applying a Particle Swarm Optimization (PSO) technique as a technique for minimizing the cost function, a radar image compensated for fluctuation is finally acquired (S120).

)는, 수학식 4로 표현 가능하다.Entropy (

) Can be expressed by equation (4).

&Quot; (4) "

은, 아래 수학식 5로 정의되는 요동 보상된 레이더 신호에 대하여, 각 버스트를 IFFT하여 획득 가능한 거리측면도에 해당된다. here,

Is a distance side view that can be obtained by IFFT each burst with respect to the oscillation compensated radar signal defined by Equation 5 below.

[Equation 5]

)로서, 속도(

)와 가속도(

) 성분에 대한 추정치(

,

)를 포함하여 구성된다. 다시 말해서, 수학식 5는 추정된 속도와 가속도 성분이 각각

과

라고 가정한 경우, 요동 보상된 레이더 신호의 수식을 나타낸다.Equation 5 above is a shake compensated radar signal (

, Speed (

) And acceleration (

) Estimate for the component (

,

It is configured to include). In other words, Equation 5 indicates that the estimated velocity and acceleration components are respectively

and

In this case, the equation of the oscillation compensated radar signal is shown.

In summary, the present invention utilizes the one-dimensional entropy minimization technique of Equation 4 above to compensate for the movement of the target in the burst using radar line speed and acceleration. Here, the entropy is used as a cost function representing the degree of focus of the fluctuation-compensated distance side view. The lower the entropy, the more focused the distance profile is.

)와 가속도(

) 성분을 추정하여, 상기 요동 보상된 레이더 영상을 최종 취득하게 된다. 물론, 이렇게 요동 보상된 레이더 영상을 취득한 이후에는 거리 압축 데이터(Range Compressed Data)를 얻을 수 있다(S130). 여기서, 상기 비용함수의 최소화를 위해 이용되는 PSO 기법에 관해서는 추후에 보다 상세히 설명하기로 한다.As described above, the present invention provides a method of estimating the speed and acceleration components in the visual direction in each burst of the radar signal while acquiring the oscillation-compensated radar image, but using a speed function that minimizes the cost function according to the entropy (

) And acceleration (

) Is estimated, and finally the oscillation compensated radar image is acquired. Of course, after the oscillation compensated radar image is acquired, range compressed data may be obtained (S130). Here, the PSO technique used to minimize the cost function will be described in more detail later.

) 및 가속도(

)를 보상할 경우, 각 버스트 내의 산란원들은 펄스별로 동일한 위치에 있게 되어, IFFT를 수행할 경우에 초점이 맞추어진 거리측면도를 얻을 수 있다. 그러나, 수학식 5의

으로 인하여 여전히 각 버스트별로 위상오차가 생기게 된다. On the other hand, referring to Equation 5, the estimated speed for the radar image (

) And acceleration (

), The scattering sources in each burst are in the same position for each pulse, so that a focused distance profile can be obtained when performing IFFT. However, in equation (5)

Due to this, there is still a phase error for each burst.

Therefore, the distance profiles computed through the IFFT of the bursts must be aligned using a range alignment algorithm. In other words, after properly compensating the fluctuation of the target within each burst, the distance side views are not aligned due to the difference in the initial position of each burst.

개의 버스트별 초기위치의 차이로 인해 정렬되어 있지 않은 거리측면도를, 1차원 엔트로피 최소화를 통해 강제 정렬하는 거리 정렬(Range Alignment)을 수행한다(S140). 또한, 레이더장치는 상기 강제로 정렬된 버스트별 위상오차를, 2차원 엔트로피 최소화를 통해 보상하는 위상 보상(Phase Adjustment)을 수행한다(S150). 이렇게 하여 최종적으로 초점이 맞추어진 레이더 영상이 취득될 수 있다(S160). 도 4는 이상과 같은 스텝주파수 ISAR 영상형성 절차로서, S110 내지 S160의 과정은 상기 레이더장치에서 수행 가능하며, 각 버스트별 요동 보상을 거치지 않을 경우에 ISAR 영상은 펄스별 움직임으로 인하여 심하게 흐려질 수 있다.That is, as described above, after obtaining the radar image compensated for the fluctuation by minimizing the cost function, the radar device is

The distance alignment that is not aligned due to the difference in the initial positions of the two bursts is forcedly aligned by minimizing the one-dimensional entropy (S140). In addition, the radar apparatus performs phase adjustment for compensating forcibly aligned bursts of each burst by minimizing two-dimensional entropy (S150). In this way, the radar image finally focused can be obtained (S160). 4 is a step frequency ISAR image forming procedure as described above, the processes of S110 to S160 can be performed in the radar apparatus, and if the burst compensation for each burst is not performed, the ISAR image may be severely blurred due to the pulse-by-pulse movement. .

On the other hand, when the fluctuation compensation (S120) step, to minimize the entropy cost function using the equation (4), the distance side view is composed of a local minima having a large cost surface (cost surface) Since it is not easy to find the minimum of the cost function, we cannot find the solution by applying the gradient descent method which is faster than other methods.

)와 가속도(

) 성분 즉, 해를 추정하게 된다. 실제 전투상황에서의 필수요건인 계산 시간을 줄이기 위하여, 각 기법에서의 세대수를 30으로 한정한 후에 각 기법의 초점정도를 비교한 결과가 후술될 것이며, 각각의 시뮬레이션 결과 PSOI 기법이 가장 적합한 것으로 나타났다.Therefore, in the present invention, as a technique for compensating the fluctuation, by applying optimization techniques of GA (Genetic Algorithm), PSO, and Particle Swarm Optimization with an Island model (PSOI), the speed to minimize entropy (

) And acceleration (

) Component, or solution. In order to reduce the calculation time, which is an essential requirement in the actual combat situation, the result of comparing the focusing degree of each technique after limiting the number of generations in each technique to 30 will be described later, and each simulation result shows that the PSOI technique is most suitable. .

Hereinafter, each technique applicable to minimizing the cost function of the present invention will be described. First, GA is an optimization technique using the concept of natural selection and evolution. GA is very effective at solving complex, multi-modal optimization problems.

와

에 대한 벡터)를 국부최적(Partical Best;pbest) 및 전역최적(Global Best; gbest)의 방향으로 변화시켜 상기 비용함수를 최소화한다. 여기서, 단계별 각 파티클의 속도벡터

를 수정하는 것은 아래 수학식 6과 같이 행해진다. 물론

와

에 대한 벡터는

에 속한다.PSO is a stochastic optimization technique based on the social behavior of a flock of birds or flocks of fish. Referring to FIG. 5, the initial random solutions called particles in the PSO technique, each velocity vector (

Wow

The cost function can be minimized by varying the vector of < RTI ID = 0.0 >< / RTI > Where the velocity vector of each particle

Is modified as in Equation 6 below. sure

Wow

Vector for

Belongs to.

&Quot; (6) "

는

라는 파티클이 지금까지 탐색한 것 중 최적의 해이고,

는 전체 파티클이 지금까지 탐색한 것 중 발견한 최적의 해이다.

는 세대번호(generation number)이고,

는 0과 1 사이에 균일분포를 가지는 랜덤한 번호이다. 이와 같이,

번째 속도벡터를 계산한 후 이를

번째 파티클

에 더함으로써,

를

와

로 움직인다. 이러한 비용함수의 최소화를 위한 최적 파티클 업데이트 및 이동 과정은 시스템이 수렴될 때까지 수행한다. 이상 상술한 PSO는 알고리즘이 간단하고, 계산 시간이 짧은 이점이 있을 뿐만 아니라, 타 알고리즘에 비해서 성능이 뛰어나기 때문에 최적화 과정에 매우 유리하다.here,

Is

Is the best year we've ever explored,

Is the best solution the whole particle has ever found.

Is the generation number,

Is a random number with a uniform distribution between 0 and 1. like this,

After calculating the second velocity vector

First particle

By adding to

To

Wow

Move to The optimal particle update and movement process to minimize this cost function is performed until the system converges. The above-described PSO is advantageous in the optimization process because the algorithm is simple, the calculation time is short, and the performance is superior to other algorithms.

On the other hand, the present invention, in order to further improve the performance of the fluctuation compensation, as a technique for minimizing the cost function, the fluctuation by applying the PSOI technique, which is a method in which a plurality of PSOs cooperate with each other to search for an optimal solution Compensation is possible. FIG. 6 illustrates an island model PSO (PSOI) that mixes three independent PSOs instead of a single PSO as an example. In other words, for a PSOI containing these three PSOs, the best performance is achieved by removing the two poorest particles in each PSO for each generation and receiving one of the best performing particles from the other two PSO populations. They work together to find a solution. Of course, to search for a more optimal solution, the PSOI may be configured using more PSO populations, but the PSO population may be limited to three or a certain number in order to reduce the computation time.

Hereinafter, simulation results for each technique (GA, PSO, PSOI) of fluctuation compensation for minimizing entropy will be described with reference to FIGS. 7 to 13. These results are analyzed by applying GA, PSO, and PSOI techniques to compensate for the pulse-to-pulse movement due to the target movement during step frequency ISAR image formation. The simulation used measurement data from a Boeing 737 aircraft and a target composed of point scatterers. As a result of the simulation, PSOI was best suited for the present invention and worked effectively on the measured ISAR data.

는 편의상 0으로 설정하였다. 그리고, 도 7의 (b)는 등방성(isotropic) 산란원으로 구성된 표적을 나타낸다. 편의상 표적의 모든 산란원의 크기는 2로 두었다.First, the results for a target composed of a point scatterer will be described. FIG. 7 (a) shows the maneuvering state of the target used on the x and y planes based on the radar apparatus for transmitting and receiving the step frequency radar signal. here,

Is set to 0 for convenience. And (b) of FIG. 7 shows the target comprised from the isotropic scattering source. For convenience, the size of all the scattering sources of the target was left at 2.

Table 1 shows the radar and motion variables used in the simulation, and Table 2 shows the variables used for GA, PSO and PSOI.

TABLE 1

TABLE 2

In Table 2, to set the variables in Table 2, the variables were selected that gave the best average result after five simulations with each variable being changed. In order to reduce the computation time, which is the most important factor in the actual battle situation, the population size and generation number in each algorithm are limited to 30. In the case of PSOI, three PSOs were used, and generation counts were distributed to 10 for each PSO.

8 shows the cost surface of the first distance side view of the target. As described above, it can be seen that the cost surface is made up of many local minimums. Each algorithm was performed 100 times at the same random position each time, and the ranges of speed and acceleration were set to 100 to 500 m / s and -10 to 70 m / s ² , respectively.

FIG. 9 (a) shows an evolution curve in which each algorithm focuses on the first distance profile for 30 generations, and FIG. 9 (b) shows distance profile compensated using each algorithm. Indicates. Table 3 below shows the mean value and standard deviation of the minimum entropy for each algorithm.

[Table 3]

Referring to Figure 9, the average of the minimum entropy in each algorithm decreases as the number of generations increases. Referring to Table 3, it can be seen that the entropy mean and standard deviation of the PSOI method are the smallest compared to the other two methods. PSO is better in accuracy than GA but slightly larger than GA in standard deviation.

In the case of PSO and PSOI calculations, each global and local optimal must be found, so it takes longer than GA to evolve the same generation. However, the GA has a disadvantage in that more generations are required in order to obtain a result similar to PSO or PSOI, which requires an increase in computation time.

10 to 11 show ISAR images without oscillation compensation and ISAR images focused using each technique. Table 4 shows the average two-dimensional entropy values and standard deviations of the images obtained after 100 independent simulations.

[Table 4]

)는 수학식 7과 같이 나타낼 수 있다.2-D entropy (

) May be expressed as in Equation 7.

[Equation 7]

According to this two-dimensional entropy standard, the image with the least entropy is the most focused image. As in the case of the 1-dimensional case, the calculation time was longer, but the performance of PSOI is the best in the focusing degree of the image. This is because ISAR image is an extension of one-dimensional distance profile. In order for GA to achieve this same result, it needs to evolve over more generations, leading to an increase in computation time. As described above, according to the simulation results using the scattering source, the performance of the PSOI was found to be the best.

와

을 다음의 수학식 8에 대입하였다.Next, a simulation result using actual measured ISAR data will be described. To this end, the performance of the PSOI is demonstrated using ISAR data measured from a real Boeing 737 aircraft. The measured data was obtained using the chirp signal. Since the bandwidth in the chirp signal is 100 MHz, the corresponding distance resolution is 1.5 m. In order to convert the step frequency data simulated earlier, Fourier transform of complex values of the distance profile is applied to the data of each frequency domain.

Wow

Was substituted into Equation 8 below.

[Equation 8]

은 거리측면도의

번째 주파수 영역 데이터를 나타내며,

은 이

에 요동 성분이 추가된 것이다. 모든 다른 변수들은 수학식 2,3과 동일하다.here,

Is the distance

The second frequency domain data,

Is this

The rocking component is added to the. All other variables are the same as in Equations 2 and 3.

12 (a) shows an original ISAR image obtained by using a chirp signal, and FIGS. 12 (b) and 13 show images before and after performing rock compensation at a step frequency. 12 (a) and 13, it can be seen that the PSOI technique performs oscillation compensation very effectively even in actual measurement data.

As described above, the present invention proposes an algorithm for compensating the phase error due to the movement of the target between pulses in the burst when forming the step frequency ISAR image, and the proposed algorithm using the target and the actual measurement data composed of scattering sources. The efficiency of each could be confirmed. In addition, when the phase error for each pulse is not compensated for, the radar image was also heavily blurred. In the case of the optimization algorithm used in the proposed method, the PSOI shows the best results in terms of focus.

Although the present invention has been described with reference to the embodiments shown in the drawings, these are merely exemplary and will be understood by those skilled in the art that various modifications and equivalent other embodiments are possible. Therefore, the true technical protection scope of the present invention will be defined by the technical spirit of the appended claims.

1 is a configuration diagram of a chirp pulse and step frequency radar,

2 is a distance side view of the chirp pulse and step frequency radar of FIG.

3 is a configuration diagram of a step frequency radar signal model used in the present invention;

4 is a schematic flowchart of a fluctuation compensation method of a step frequency radar image using a PSO according to an embodiment of the present invention;

5 is an explanatory diagram of a PSO technique applied to FIG. 4;

6 is a configuration diagram of a PSOI technique that is an application of the PSO technique of FIG.

7 is an illustration of the target situation and target used for ISAR imaging;

8 is a cost surface of the first distance side view according to FIG.

FIG. 9 is a distance side view oscillated using an evolutionary curve GA focusing on the first distance side view according to FIG. 7 and each algorithm; FIG.

10 to 11 are images when the oscillation compensation is not performed, ISAR images after the oscillation compensation through each technique,

12 to 13 are comparisons of images before and after shaking compensation using measured data.

Claims (6)

After receiving the step frequency radar signal reflected from the target transmitted to the target side to obtain a radar image, In order to correct the phase error according to the fluctuation of the target, the radar is compensated for the fluctuation by applying the particle swarm optimization (PSO) technique as a technique for setting the entropy of the radar image as a cost function and minimizing the cost function. A fluctuation compensation method of a step frequency radar image using PSO for finally obtaining an image. The method according to claim 1, wherein the step frequency radar signal,

Of pulses

A method for compensating for shaking of a step frequency radar image using PSO, comprising 4 bursts. The method according to claim 2, wherein upon acquisition of the oscillation compensated radar image, Estimating the velocity and acceleration components in the visual direction in each burst of the radar signal, but estimating the velocity and acceleration components to minimize the cost function and finally obtaining the oscillation compensated radar image. Fluctuation compensation method of frequency radar image. The method according to claim 2 or 3, wherein the entropy (

),

Defined as above

Is a distance side view obtained by IFFT each fluctuation-compensated burst, the fluctuation compensation method of a step frequency radar image using PSO. The method according to claim 2, wherein after obtaining the oscillation compensated radar image using the minimization of the cost function, remind

Performing a distance alignment to force-align the distance side views which are not aligned due to the difference of initial positions for each burst, and performing phase adjustment to compensate the phase error for each forced burst. Fluctuation compensation method of step frequency radar image using PSO. The method according to claim 1, As a technique for minimizing the cost function, the fluctuation compensation can be performed by applying a Particle Swarm Optimization with an Island model (PSOI) technique, in which a plurality of PSOs cooperate with each other to search for an optimal solution. Fluctuation compensation method of step frequency radar image using PSO.

Images