Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group

Bi-frequency correlation properties of the scattered intensity from dielectric rough surfaces

Open Access Open Access

Abstract

The analytical expression for the Bi-frequency correlation function of the intensity scattered from two-dimensional dielectric randomly rough surfaces obeying Gaussian distribution are presented based on the scalar Kirchhoff approximation theory with the root-mean-square (rms) slope of the surface less than 0.25 and the Gaussian moment theorem. The results show that the bi-frequency correlation properties of the scattered intensity closely depend on the incident and scattered conditions as well as on the statistical parameters and complex refractive index of the surface. Especially, the correlation function mainly comes from the specular direction, and the coherence bandwidth and the function decrease with the increase of the roughness of the rough surface. In addition, comparing with the real part, the imagery of the complex refractive index has a greater impact on the bi-frequency correlation function.

©2012 Optical Society of America

1. Introduction

The problem of the scatterings of electromagnetic waves and optics from randomly rough surfaces has always been a very important topic, and has very widely scientific and technical applications, such as in radar, SAR remote sensing, surface detection and target recognition [15]. In many radio and optical measurements, regardless of the radar waves or the lasers emitting pulses with some bandwidth, it is necessary to study the correlation of the scattered fields with different frequency. Ishimaru employed the two-frequency mutual coherence function to study pulse scattering from rough surfaces and discussed the pulse broadening and the enhanced backscattering effect [6]. Schertler and George derived the formulas of the two-frequency mutual correlation function of the backscattering from roughened sphere and roughened disk [7, 8]. Chen et al derived the two-frequency mutual coherent function to investigate the pulse scattering properties of the pulse plane wave and pulse beam from randomly rough surfaces [9]. The two-frequency mutual coherence function of the scattering from arbitrarily shaped rough objects were obtained and the numerical results for rough spheres and cylinders were given to analyze the dependence of the function on the shape and the size of the objects and on the roughness of the surface [10].

In practice, the intensity not the scattered field is the directly detected parameter by experiments; therefore, the correlation properties of the scattering intensity are of more considerable interest. The solution for the fourth moment equation of waves in random media was given by Xu et al [11]. The frequency spectrum of the intensity fluctuations of the scattered field was used to measure the parameters of vibrations and the surface roughness, and the experimental results were presented [12]. The fourth order moment statistical characteristics of the wave scattering from random rough surfaces were derived and the two-frequency mutual coherence function of the scattered intensity was investigated numerically [13]. The correlation between intensity fluctuations of light scattered from a quasi-homogeneous random media was analytically derived, and its dependence on spatial Fourier transforms of both the intensity and degree of spatial correlation of scattering potentials were analyzed [14]. The expressions for the fourth-order moment of a random field incident upon deterministic and random scatterers were derived which can serve as a rigorous analytic prediction of a scattered field by scatterers and demonstrated how the intensity–intensity correlations were affected at various points in the far field [15]. Most of the discussions above simplified the correlation function of the intensity as the square of the module of the correlation function of the scattered field, since the surface considered was rougher than the incident wavelength. With Gaussian moment theorem, the correlation properties of the intensity scattered from one-dimensional and two-dimensional randomly conducting rough surfaces with RMS roughness smaller than the wavelength were discussed [16, 17] in detail. However, during the investigations the influences of the slopes and the dielectric property of the surface on the intensity correlation function were not considered for simplicity.

Based on the scalar Kirchhoff approximation theory, the bi-frequency correlation function of the scattered intensity from two-dimensional dielectric randomly rough surface with a root-mean -square slope smaller than 0.25 is investigated in this paper. Assuming the distribution of the rough surface to be Gaussian, the expression for the bi-frequency correlation function of the scattered intensity is obtained, and then the results and the influence factors are presented. It is noted that the bi-frequency correlation function of the scattered intensity closely dependent on the roughness of the surface mainly comes from the specular direction when the surface is weakly fluctuating.

2. Bi-frequency correlation function of the scattered intensity from dielectric rough surfaces

When the surface discussed is weakly rough that the roughness is smaller than the incident wavelength, the multiple scattering from different area of the rough surface and the shadowing effect can be omitted. The rms slope of the surface smaller than 0.25, according to the scalar Kirchhoff approximation theory, the scattered field from two-dimensional randomly dielectric rough surface can be expressed as [18]

Epqs(f)=E0K(f)p(x,y)Upqexp[i(kski)r]dS
whereK(f)=ikexp(ikR)/4πR, and k=2π/λ=2πf/cis the wave-number in free space,λis the incident wavelength,fis the incident frequency,cis the light velocity. The aperture function p(x,y)=exp[(x2+y2)/D2] is introduced to expand the integral domain to infinite, andDis its dimension. Upq=a0+a1Zx+a2Zy, wherea0,a1,a2are polarization coefficients whose expressions are in reference [18], andZx,Zyrepresent the slope of the surface alonge^x,e^ydirections at the point r, respectively.r=xe^x+ye^y+h(x,y)e^z,h(x,y)is the fluctuating function of the surface, and ks=k(sinθscosφs,sinθscosφs,cosθs), ki=k(sinθicosφi,sinθicosφi,cosθi)are the scattered and incident wave vectors, where (θs,φs)and(θi,φi)are the scattered and incident directions, respectively.

The bi-frequency correlation function of the intensity scattered from randomly rough surfaces is defined as

C12=I(f1)I(f2)I(f1)I(f2)=|Epqs(f1)|2|Epqs(f2)|2|Epqs(f1)|2|Epqs(f2)|2 (2)

The rms slope of the surface discussed in this paper is smaller than 0.25 and its rms roughness is less than the incident wavelength, the surface is weak rough surface, therefore, the scattering composes coherent and incoherent parts, that is

Epqs(f)=ΔEpqs(f)+Epqs(f)

The bi-frequency correlation function of the scattered intensity can be written as follow [17]

C12=|Ipq1(f1,f2)|2+|Ipq2(f1,f2)|22Epq1s2Epq2s2
whereIpq1(f1,f2)=Epq1sEpq2sandIpq2(f1,f2)=Epq1sEpq2s.

2.1 Solution for the termIpq1(f1,f2)

According to Eq. (1), the termIpq1(f1,f2)is

Ipq1=Upq(f1)Upq*(f2)exp[i(V1r1V2r2)]dS1dS2
neglecting the unnecessary factor before the integral, andV1,2=(kski)1,2. For convenience, let the vectorsVandrbeV=V+Vze^z,r=r+h(r)e^z. And the termUpq(f1)Upq*(f2)can be written as

Upq(f1)Upq*(f2)=a01a02*+a11a02*Zx1+a01a12*Zx2+a21a02*Zy1+a01a22*Zy2

The second number in the subscript of the polarization coefficient represents the frequency. And here the second terms of the slopes are omitted, which because when the maximum radius of curvature of the surface is much greater than the incident wavelength, the second terms of the slope will be much smaller than the first terms.

From Eq. (6), it can be seen that Ipq1(f1,f2)can be decomposed into two terms: non-slope term and slope term,

Ipq1(f1,f2)=Ipq10(f1,f2)+Ipq1s(f1,f2)

The non-slope term is

Ipq10=a01a02*exp[i(Vz1h1Vz2h2)]exp[i(V1r1V2r2)]dS1dS2

The slope term includes x-slope term and y-slope term

Ipq1s(f1,f2)=Ipq1x(f1,f2)+Ipq1y(f1,f2)
where
Ipq1x=a11a02*Ipq1x1+a01a12*Ipq1x2
Ipq1y=a21a02*Ipq1y1+a01a22*Ipq1y2
with
Ipq1xj=Zxjexp[i(Vz1h1Vz2h2)]exp[i(V1r1V2r2)]dS1dS2
Ipq1yj=Zyjexp[i(Vz1h1Vz2h2)]exp[i(V1r1V2r2)]dS1dS2
wherej=1,2.

The rough surface is assumed to be Gaussian, then the joint-characteristic function has the form

F(Vz1,Vz2,ρ)=exp[i(Vz1h1Vz2h2)]=exp[(β11+β22)/2]exp[β12ρ(r1r2)]
whereβij=VziVzjδ2(i,j=1,2), and ρ(r1r2)=exp[|r1r2|2/lc2]is the correlation coefficient of the rough surface,δand lcare the rms roughness and the correlation length, respectively. For a Gaussian randomly rough surface, there are the relationships as follow [18]
Zxjexp[i(Vz1h1Vz2h2)]=iVz(3j)δ2ρxdF(Vz1,Vz2,ρ)
Zyjexp[i(Vz1h1Vz2h2)]=iVz(3j)δ2ρydF(Vz1,Vz2,ρ)
wherexd=x1x2,yd=y1y2.

Making the following change of variables

rc=(r1+r2)/2rd=r1r2Vc=(V1+V2)/2Vd=V1V2

Substituting Eqs. (14)and(17) into Eq. (8) gives the non-slope terms as

Ipq10=a01a02*βdrcdrdexp(2|rc|2D2)exp(|rd|22D2)exp(β12ρ)exp[i(Vdrc+Vcrd)]
withβ=exp[(β11+β22)/2].

To evaluate the integral, we must expand the termexp(β12ρ)into Taylor series

exp[β12ρ(rd)]=n=0(β12)nn!ρn(rd)

Then the solution is

Ipq10=a01a02*π2D4lc2βexp[D2|Vd|28]n=0(β12)nn!(lc2+2nD2)exp(D2lc2|Vc|22lc2+4nD2)

To solve the integrals in Eqs. (12) and (13), some new variables are introduced

xd=ξcosαyd=ξsinα

The partial derivatives of the functionρbecome

ρxd=ρξcosαρyd=ρξsinα

Inserting Eqs. (15), (17), (21) and (22) in Eq. (12) and carrying out the integration overrc, yields

Ipq1xj=iVz(3j)πD2δ2β2exp[D2|Vd|28]0ξdξexp(β12ρ)exp(ξ22D2)ρξΓ(ξ)
where
Γ(ξ)=02πdαcosαexp[iξ|Vc|cos(αχ)]
andχ=arctan(Vcy/Vcx).

With the complex exponential term in Eq. (23) which can be expressed as [19]

exp[ibcos(αχ)]=n=inJn(b)exp[in(αχ)]
and the integral relationships [20]
02πdαcosαexp(inα)={πn=±10n±1
02πdαsinαexp(inα)={±iπn=±10n±1
andJn=(1)nJn, then the functionΓ(ξ)has the solution
Γ(ξ)=i2πJ1(ξ|Vc|)cosχ
whereJn(...)is the nth-order Bessel function.

Inserting Eq. (28) into Eq. (23) and utilizing Eq. (19), the final solution for Eq. (12) can be attained

Ipq1xj=2π2D6Vz(3j)δ2βVcxexp(D2|Vd|2/8)×n=0lc2β12nn![lc2+2(1+n)D2]2exp{D2lc2|Vc|22[lc2+2(1+n)D2]}

In the same way, the solution of Eq. (13) can be obtained as

Ipq1yj=2π2D6Vz(3j)δ2βVcyexp(D2|Vd|2/8)×n=0lc2β12nn![lc2+2(1+n)D2]2exp{D2lc2|Vc|22[lc2+2(1+n)D2]}

Then substituting Eq. (29) and Eq. (30) into Eq. (10) and Eq. (11), respectively, the solution for the slope termIpq1sis

Ipq1s=[(a11a02*Vz2+a01a12*Vz1)Vcx+(a21a02*Vz2+a01a22*Vz1)Vcy]2π2D6δ2β×exp[D2|Vd|28]n=0lc2β12nn![lc2+2(1+n)D2]2exp{D2lc2|Vc|22[lc2+2(1+n)D2]}
With Eqs. (20)and(31), the analytical expression for the termIpq1(f1,f2)can be achieved.

2.2 Solution for the termIpq2(f1,f2)

Similarly, from Eq. (1) the termIpq2(f1,f2) can be written as

Ipq2=Upq1Upq2exp[i(V1r1+V2r2)]dS1dS2
where

Upq1Upq2=a01a02+a11a02Zx1+a01a12Zx2+a21a02Zy1+a01a22Zy2

Also Ipq2(f1,f2)can be expressed by the non-slope term and the slope term

Ipq2(f1,f2)=Ipq20(f1,f2)+Ipq2s(f1,f2)
and

Ipq20=a01a02exp[i(V1r1+V2r2)]exp[i(Vz1h1+Vz2h2)]dr1dr2

The slope term Ipq2s(f1,f2) is then

Ipq2s(f1,f2)=Ipq2x(f1,f2)+Ipq2y(f1,f2)
where
Ipq2x=a11a02Ipq2x1+a01a12Ipq2x2
Ipq2y=a21a02Ipq2y1+a01a22Ipq2y2
and

Ipq2xj=Zxjexp[i(Vz1h1+Vz2h2)]exp[i(V1r1+V2r2)]dr1dr2
Ipq2yj=Zyjexp[i(Vz1h1+Vz2h2)]exp[i(V1r1+V2r2)]dr1dr2

Here the joint-characteristic function is

F(Vz1,Vz2,ρ)=exp[i(Vz1h1+Vz2h2)]=βexp(β12ρ)
and

Zxjexp[i(Vz1h1+Vz2h2)]=(1)3jiVz(3j)δ2ρξcosαF(Vz1,Vz2,ρ)
Zyjexp[i(Vz1h1+Vz2h2)]=(1)3jiVz(3j)δ2ρξsinαF(Vz1,Vz2,ρ)

By Eqs. (17)and(21), and finishing the integration overrc, Eqs. (35) and (39)-(40) can be transformed as follow

Ipq20=a01a02πβD22exp(D2|Vc|22)0ξdξexp(ξ22D2)exp(β12ρ)Η(ξ)
Ipq2xj=(1)3jiδ2Vz(3j)πβD22exp(D2|Vc|22)0ξdξexp(ξ22D2)ρξexp(β12ρ)Λ(ξ)
Ipq2yj=(1)3jiδ2Vz(3j)πβD22exp(D2|Vc|22)0ξdξexp(ξ22D2)ρξexp(β12ρ)Φ(ξ)
where
Η(ξ)=02πdαexp[iξ|Vd|2cos(αε)]
Λ(ξ)=02πdαcosαexp[iξ|Vd|2cos(αε)]
Φ(ξ)=02πdαsinαexp[iξ|Vd|2cos(αε)]
withε=arctan(Vyd/Vxd).

By Eqs. (25)-(27), The functionsΗ(ξ),Λ(ξ)andΦ(ξ)have the solutions

Η(ξ)=2πJ0[ξ2|Vd|]
Λ(ξ)=i2πJ1(ξ2|Vd|)cosε
Φ(ξ)=i2πJ1(ξ2|Vd|)sinε

Noting that in Eqs. (44)-(46), because of the termexp[β12ρ(ξ)], it is difficult to get the solutions analytically, a more simple approximation should be taken which we have discussed in our paper [16] before. Using the Eq. (16) in the reference [16], and inserting Eqs. (50)-(52)into Eqs. (44)-(46), respectively, we can get

Ipq20=π2D2a01a02βexp(D2|Vc|22){exp(β12)0ξ0ξdξexp(ξ22D2)×J0[ξ2|Vd|]+ξ0ξdξexp(ξ22D2)J0[ξ2|Vd|]}
Ipq2xj=(1)4jVz(3j)π2D2δ2βcosεexp(D2|Vc|22){exp(β12)0ξ0ξdξ×exp(ξ22D2)ρξJ1[ξ2|Vd|]+ξ0ξdξexp(ξ22D2)ρξJ1[ξ2|Vd|]}
Ipq2yj=(1)4jVz(3j)π2D2δ2βsinεexp(D2|Vc|22){exp(β12)0ξ0ξdξ×exp(ξ22D2)ρξJ1[ξ2|Vd|]+ξ0ξdξexp(ξ22D2)ρξJ1[ξ2|Vd|]}
where

ξ0=lc{ln(β12ln2ln[1+exp(β12)])}1/2

Since the Bessel functions of integer orderJn(z) can be expressed explicitly as [21]

Jn(z)=k=0(1)kz2k+n22k+nk!(k+n)!

And there are the relationships [20]

γ(a,x)=0xetta1dtΓ(a,x)=xetta1dt[Rea>0]

The final solutions forIpq20andIpq2scan be given by

Ipq20=π2D4a01a02βexp(D2|Vc|22)k=0(1)k(|Vd|D)2k23k(k!)2×[exp(β12)γ(k+1,ξ022D2)+Γ(k+1,ξ022D2)]
Ipq2s=[(a11a02Vz2a01a12Vz1)Vdx+(a21a02Vz2a01a22Vz1)Vdy]π2D6δ2β×exp(D2|Vc|22)k=0(1)k|Vd|2kD2klc2k+223kk!(k+1)!(lc2+2D2)k+2[exp(β12)γ(k+2,x0)+Γ(k+2,x0)]
withx0=ξ02/2D2+ξ02/lc2.

2.3 Mean scattered fieldEpqjs(fj)

At last, the mean scattered fieldEpqjs(fj)should be derived. By Eq. (1), Epqjs(fj) can be written as

Epqjs=(a0j+a1jZx+a2jZy)exp(iVzjh)exp(iVjr)dS

The term exp(iVzjh) in the equation above is the characteristic function of the Gaussian random variable

exp[iVzjh(r)]=exp(Vzj2δ2/2)=exp(βjj/2)
and by reference [18]
Zxexp[iVzjh(r)]=Zyexp[iVzjh(r)]=0
The mean scattered field is then

Epqjs=πD2a0jexp(βjj/2)exp(D2|Vj|2/4)

Thus every term needed has been derived, inserting these solutions in Eq. (4), the analytical expression for the bi-frequency correlation function of the scattered intensity from two- dimensional randomly dielectric weak rough surface can then be obtained.

3. Numerical results and analyses

In this section, the bi-frequency correlation function of the scattered intensity will be calculated numerically to analyze its bi-frequency correlation properties and the influencing factors. In the calculation, let the center frequency bef0=(f1+f2)/2and the frequency difference befd=f1f2, and the medium be assumed non-dispersive or in the extent of the bandwidth the refractive index have hardly change, thus for the frequency differencefd, the polarization coefficientsa0j,a1j,a2j are constant. From the formulas derived above, it can be seen that the bi-frequency correlation function C12of the scattered intensity from two-dimensional dielectric rough surface with a root-mean-square slope smaller than 0.25 and rms roughness less than the incident wavelength is closely dependent on the incident and scattered conditions as well as on the statistical parameters and refractive index of the surface.

Taking the wavelength is1.06μmand the aperture dimension is13.3lc as well as the refractive index isn=2.43+i10.7, Figs. 1 -4 are the variation of he bi-frequency correlation functionC12of the scattered intensity with the scattering angleθsand the frequency differencefd. In Figs. 1-3 , θi=φi=φs=0°,lc=5.89μm, the rms roughnessδincreases sequentially. Comparing with Fig. 3, in Fig. 4 the incident angle is10°, δis invariable, δ=0.9μm,the correlation length is increased to be7.5μm.

 figure: Fig. 1

Fig. 1 Bi-frequency correlation functionC12versus scattering angle and frequency difference with the parametersδ=0.6μmandlc=5.89μm

Download Full Size | PDF

 figure: Fig. 4

Fig. 4 Bi-frequency correlation function C12versus scattering angle and frequency difference with the parametersθi=10°,δ=0.9μmandlc=7.5μm

Download Full Size | PDF

 figure: Fig. 3

Fig. 3 Bi-frequency correlation function C12versus scattering angle and frequency difference with the parametersδ=0.9μmandlc=5.89μm

Download Full Size | PDF

 figure: Fig. 2

Fig. 2 Bi-frequency correlation function C12versus scattering angle and frequency difference with the parametersδ=0.8μmandlc=5.89μm

Download Full Size | PDF

It is shown that the contribution for the functionC12mainly comes from the specular direction while the values in other directions can be neglected. With a smaller rms slopes=2δ/lc, the decrease of the function C12versus the frequency differencefdis much more slowly that the correlation bandwidth is bigger and the profile near the specular is sharper. While with an increases, the values of the function C12 in specular direction decreases dramatically and others increase relatively. It is explained that when the rms slope is small, the scattering mode is similar to the plane scattering which scattered power only occurs in specular direction, while the surface becomes more roughness, the non-specular scattering enhances and the specular scattering weakens. In addition, the increase of rms height results in a narrower correlation bandwidth.

θi=θs=φi=0°andlc=5.89μm, with different roughnessδ, Figs. 5 -6 and Figs. 7 -8 are the variation of the bi-frequency correlation function C12 against the scattering azimuth angleφs and the frequency difference fdunder HH- and VH-polarization, respectively. Under both the polarizations, the functionC12decreases monotonously with the increase of the frequency differencefd and the roughnessδ. A bigger roughnessδresults in a smaller correlation bandwidth. However, in HH-polarization the variable tendency of the functionC12with the increase of the azimuth angleφs is completely different to that in VH-polarization.

 figure: Fig. 5

Fig. 5 Bi-frequency correlation function C12 versus scattering azimuth angle and frequency difference with the parametersδ=0.6μm andlc=5.89μm, HH-polarization

Download Full Size | PDF

 figure: Fig. 6

Fig. 6 Bi-frequency correlation function C12 versus scattering azimuth angle and frequency difference with the parametersδ=0.8μm andlc=5.89μm, HH-polarization

Download Full Size | PDF

 figure: Fig. 7

Fig. 7 Bi-frequency correlation functionC12versus scattering azimuth angle and frequency difference with the parameters δ=0.6μmandlc=5.89μm, VH-polarization

Download Full Size | PDF

 figure: Fig. 8

Fig. 8 Bi-frequency correlation function C12 versus scattering azimuth angle and frequency difference with δ=0.8μm andlc=5.89μm, VH-polarization

Download Full Size | PDF

To illustrate the influence of the dielectric property of the rough surface on the bi-frequency functionC12, Fig. 9 gives the profiles of the functionC12versus the frequency differencefdwith different roughness and refractive indexes. It is can be seen easily that with a complex refractive index, the smaller the roughness, the bigger the functionC12, while with a same roughness, the refractive index has influence on the value of the functionC12not on the correlation bandwidth. Especially, the imagery of the refractive index is the main influencing factor that a smaller difference of the absolute values of the imagery of the refractive indexes results in a smaller difference of the values of the functionC12.

 figure: Fig. 9

Fig. 9 Bi-frequency correlation function C12 versus frequency difference with different refractive indexes

Download Full Size | PDF

4. Conclusion

Based on the scalar Kirchhoff approximation theory, the Bi-frequency correlation properties of the intensity scattered from two-dimensional dielectric randomly rough Gaussian surfaces assumed to be Gaussian with a root-mean-square (rms) slope of the surface less than 0.25 are investigated in this paper. The rms roughness of the surface is assumed to be smaller than the incident wavelength. By the Gaussian moment theorem, expanding the exponential and complex exponential function into series and with some mathematical simplification, the Bi-frequency correlation function of the scattered intensity considering the influences of the scattering from the slope of the surface has been derived in detail and calculated. The numerical results show that the incident and scattered conditions as well as on the statistical parameters and complex refractive index of the surface have great impacts on the bi-frequency correlation properties of the scattered intensity. Especially, the correlation function mainly comes from the specular direction, and the coherence bandwidth decreases with the increase of the roughness of the rough surface. In HH-polarization and VH-polarization, the functionC12versus with the scattering azimuth angle have different varying tendency. In addition, the refractive index can also affect the value of the functionC12 but not on the correlation bandwidth. Comparing with the real part, the imagery of the complex refractive index has a greater influence on the bi-frequency correlation function that the larger difference of the absolute values of the imagery of the refractive indexes, the larger difference of the values of the functionC12. The work performed in this paper provide a much better knowledge about the scattering properties of dielectric rough surfaces to investigate the scattering problem and the identification of three-dimensional rough dielectric objects which we will discuss in future.

Acknowledgments

The authors gratefully acknowledge support from the National Natural Science Foundation of China under Grant No. 61172031 and from the Fundamental Research Funds for the Central Universities under Grant No. K50510070009.

References and links

1. A. Ishimaru, Wave Propagation and Scattering in Random Media (Academic Press, 1978).

2. P. Beckmann and A. Spizzichino, The Scattering of Electromagnetic Waves from Rough Surfaces (Pergamon, 1963).

3. E. Bahar and S. Chakrabarti, “Scattering and depolarization by large conducting spheres with rough surfaces,” Appl. Opt. 24(12), 1820–1825 (1985). [CrossRef]   [PubMed]  

4. E. Bahar and M. A. Fitzwater, “Scattering and depolarization by conducting cylinders with rough surfaces,” Appl. Opt. 25(11), 1826–1832 (1986). [CrossRef]   [PubMed]  

5. W. Zhensen and C. Suomin, “Bistatic scattering by arbitrarily shaped objects with rough surface at optical and infrared frequencies,” Int. J. Infrared Millim. Waves 13(4), 537–549 (1992). [CrossRef]  

6. A. Ishimaru, L. Ailes-Sengers, P. Phu, and D. Winebrenner, “Pulse broadening and two-frequency mutual coherence function of the scattered wave from rough surfaces,” Waves Random Media 4(2), 139–148 (1994). [CrossRef]  

7. D. J. Schertler and N. George, “Backscattering cross section of a tilted, roughened disk,” J. Opt. Soc. Am. A 9(11), 2056–2066 (1992). [CrossRef]  

8. D. J. Schertler and N. George, “Backscattering cross section of a roughened sphere,” J. Opt. Soc. Am. A 11(8), 2286–2297 (1994). [CrossRef]  

9. C. Hui, W. Zhensen, and B. Lu, “Infrared laser pulse scattering from randomly rough surfaces,” Int. J. Infrared Millim. Waves 25(8), 1211–1219 (2004). [CrossRef]  

10. G. Zhang and Z. S. Wu, “Two-frequency mutual coherence function of scattering from arbitrarily shaped rough objects,” Opt. Express 19(8), 7007–7019 (2011). [CrossRef]   [PubMed]  

11. Z. W. Xu, J. Wu, Z. S. Wu, and Q. Li, “Solution for the Fourth Moment Equation of Waves in Random Continuum Under Strong Fluctuations: General Theory and Plane Wave Solution,” IEEE Trans. Antenn. Propag. 55(6), 1613–1621 (2007). [CrossRef]  

12. V. N. Bronnikov and M. M. Kalugin, “Measuring the parameters of vibrations and surface roughness, using the frequency spectrum of the intensity fluctuations of scattered radiation,” J. Opt. Tech. 76(11), 697–701 (2009). [CrossRef]  

13. M. J. Wang, Z. S. Wu, Y. L. Li, X. A. Zhang, and H. Zhang, “The fourth order moment statistical characteristic of the laser pulse scattering on random rough surface,” Acta Phys. Sin-CH ED 58, 2390–2396 (2009).

14. Y. Xin, Y. J. He, Y. R. Chen, and J. Li, “Correlation between intensity fluctuations of light scattered from a quasi-homogeneous random media,” Opt. Lett. 35(23), 4000–4002 (2010). [CrossRef]   [PubMed]  

15. H. C. Jacks and O. Korotkova, “Intensity-intensity fluctuations of stochastic fields produced upon weak scattering,” J. Opt. Soc. Am. A 28(6), 1139–1144 (2011). [CrossRef]   [PubMed]  

16. W. Zhen-Sen and Z. Geng “Intensity Correlation Function of Light Scattering from a Weakly One-Dimensional Random Rough Surface,” Chin. Phys. Lett. 26(11), 114208 (2009). [CrossRef]  

17. G. Zhang and Z. Wu, “Fluctuation correlation of the scattered intensity from two-dimensional rough surfaces,” Opt. Express 20(2), 1491–1502 (2012). [CrossRef]   [PubMed]  

18. F. T. Ulaby, R. K. Moore, and A. K. Fung, Microwave Remote Sensing, Vol. 2 (Addison-Wesley Publishing,1982).

19. C. Bourlier, “Azimuthal Harmonic Coefficients of the Microwave Backscattering from a Non-Gaussian Ocean Surface With the First-Order SSA Model,” IEEE Trans. Geosci. Rem. Sens. 42(11), 2600–2611 (2004). [CrossRef]  

20. J. S. Gradshteyn and J. M. Ryzhik, Table of Integrals, Series and Products (Academic, 1965).

21. G. N. Watson, A Treatise on the Theory of Bessel Functions (Cambridge University Press, 1958).

Cited By

Optica participates in Crossref's Cited-By Linking service. Citing articles from Optica Publishing Group journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (9)

Fig. 1
Fig. 1 Bi-frequency correlation function C 12 versus scattering angle and frequency difference with the parameters δ=0.6μm and l c =5.89μm
Fig. 4
Fig. 4 Bi-frequency correlation function C 12 versus scattering angle and frequency difference with the parameters θ i = 10 ° , δ=0.9μm and l c =7.5μm
Fig. 3
Fig. 3 Bi-frequency correlation function C 12 versus scattering angle and frequency difference with the parameters δ=0.9μm and l c =5.89μm
Fig. 2
Fig. 2 Bi-frequency correlation function C 12 versus scattering angle and frequency difference with the parameters δ=0.8μm and l c =5.89μm
Fig. 5
Fig. 5 Bi-frequency correlation function C 12 versus scattering azimuth angle and frequency difference with the parameters δ=0.6μm and l c =5.89μm , HH-polarization
Fig. 6
Fig. 6 Bi-frequency correlation function C 12 versus scattering azimuth angle and frequency difference with the parameters δ=0.8μm and l c =5.89μm , HH-polarization
Fig. 7
Fig. 7 Bi-frequency correlation function C 12 versus scattering azimuth angle and frequency difference with the parameters δ=0.6μm and l c =5.89μm , VH-polarization
Fig. 8
Fig. 8 Bi-frequency correlation function C 12 versus scattering azimuth angle and frequency difference with δ=0.8μm and l c =5.89μm , VH-polarization
Fig. 9
Fig. 9 Bi-frequency correlation function C 12 versus frequency difference with different refractive indexes

Equations (63)

Equations on this page are rendered with MathJax. Learn more.

E pq s ( f )= E 0 K( f ) p( x,y ) U pq exp[ i( k s k i )r ]dS
E pq s ( f )=Δ E pq s ( f )+ E pq s ( f )
C 12 = | I pq1 ( f 1 , f 2 ) | 2 + | I pq2 ( f 1 , f 2 ) | 2 2 E pq1 s 2 E pq2 s 2
I pq1 = U pq ( f 1 ) U pq * ( f 2 )exp[ i( V 1 r 1 V 2 r 2 ) ] d S 1 d S 2
U pq ( f 1 ) U pq * ( f 2 )= a 01 a 02 * + a 11 a 02 * Z x1 + a 01 a 12 * Z x2 + a 21 a 02 * Z y1 + a 01 a 22 * Z y2
I pq1 ( f 1 , f 2 )= I pq10 ( f 1 , f 2 )+ I pq1s ( f 1 , f 2 )
I pq10 = a 01 a 02 * exp[ i( V z1 h 1 V z2 h 2 ) ] exp[ i( V 1 r 1 V 2 r 2 ) ]d S 1 d S 2
I pq1s ( f 1 , f 2 )= I pq1x ( f 1 , f 2 )+ I pq1y ( f 1 , f 2 )
I pq1x = a 11 a 02 * I pq1x1 + a 01 a 12 * I pq1x2
I pq1y = a 21 a 02 * I pq1y1 + a 01 a 22 * I pq1y2
I pq1xj = Z xj exp[ i( V z1 h 1 V z2 h 2 ) ] exp[ i( V 1 r 1 V 2 r 2 ) ]d S 1 d S 2
I pq1yj = Z yj exp[ i( V z1 h 1 V z2 h 2 ) ] exp[ i( V 1 r 1 V 2 r 2 ) ]d S 1 d S 2
F( V z1 , V z2 ,ρ )= exp[ i( V z1 h 1 V z2 h 2 ) ] =exp[ ( β 11 + β 22 ) /2 ]exp[ β 12 ρ( r 1 r 2 ) ]
Z xj exp[ i( V z1 h 1 V z2 h 2 ) ] =i V z(3j) δ 2 ρ x d F( V z1 , V z2 ,ρ )
Z yj exp[ i( V z1 h 1 V z2 h 2 ) ] =i V z(3j) δ 2 ρ y d F( V z1 , V z2 ,ρ )
r c = ( r 1 + r 2 ) /2 r d = r 1 r 2 V c = ( V 1 + V 2 ) /2 V d = V 1 V 2
I pq10 = a 01 a 02 * β d r c d r d exp( 2 | r c | 2 D 2 )exp( | r d | 2 2 D 2 )exp( β 12 ρ )exp[ i( V d r c + V c r d ) ]
exp[ β 12 ρ( r d ) ]= n=0 ( β 12 ) n n! ρ n ( r d )
I pq10 = a 01 a 02 * π 2 D 4 l c 2 βexp[ D 2 | V d | 2 8 ] n=0 ( β 12 ) n n!( l c 2 +2n D 2 ) exp( D 2 l c 2 | V c | 2 2 l c 2 +4n D 2 )
x d =ξcosα y d =ξsinα
ρ x d = ρ ξ cosα ρ y d = ρ ξ sinα
I pq1xj =i V z(3j) π D 2 δ 2 β 2 exp[ D 2 | V d | 2 8 ] 0 ξdξexp( β 12 ρ )exp( ξ 2 2 D 2 ) ρ ξ Γ( ξ )
Γ( ξ )= 0 2π dαcosα exp[ iξ| V c |cos( αχ ) ]
exp[ ibcos( αχ ) ]= n= i n J n ( b )exp[ in( αχ ) ]
0 2π dαcosαexp( inα ) ={ π n=±1 0 n±1
0 2π dαsinαexp( inα ) ={ ±iπ n=±1 0 n±1
Γ( ξ )=i2π J 1 ( ξ| V c | )cosχ
I pq1xj =2 π 2 D 6 V z(3j) δ 2 β V cx exp( D 2 | V d | 2 /8 ) × n=0 l c 2 β 12 n n! [ l c 2 +2( 1+n ) D 2 ] 2 exp{ D 2 l c 2 | V c | 2 2[ l c 2 +2( 1+n ) D 2 ] }
I pq1yj =2 π 2 D 6 V z(3j) δ 2 β V cy exp( D 2 | V d | 2 /8 ) × n=0 l c 2 β 12 n n! [ l c 2 +2( 1+n ) D 2 ] 2 exp{ D 2 l c 2 | V c | 2 2[ l c 2 +2( 1+n ) D 2 ] }
I pq1s =[ ( a 11 a 02 * V z2 + a 01 a 12 * V z1 ) V cx +( a 21 a 02 * V z2 + a 01 a 22 * V z1 ) V cy ]2 π 2 D 6 δ 2 β ×exp[ D 2 | V d | 2 8 ] n=0 l c 2 β 12 n n! [ l c 2 +2( 1+n ) D 2 ] 2 exp{ D 2 l c 2 | V c | 2 2[ l c 2 +2( 1+n ) D 2 ] }
I pq2 = U pq1 U pq2 exp[ i( V 1 r 1 + V 2 r 2 ) ] d S 1 d S 2
U pq1 U pq2 = a 01 a 02 + a 11 a 02 Z x1 + a 01 a 12 Z x2 + a 21 a 02 Z y1 + a 01 a 22 Z y2
I pq2 ( f 1 , f 2 )= I pq20 ( f 1 , f 2 )+ I pq2s ( f 1 , f 2 )
I pq20 = a 01 a 02 exp[ i( V 1 r 1 + V 2 r 2 ) ] exp[ i( V z1 h 1 + V z2 h 2 ) ] d r 1 d r 2
I pq2s ( f 1 , f 2 )= I pq2x ( f 1 , f 2 )+ I pq2y ( f 1 , f 2 )
I pq2x = a 11 a 02 I pq2x1 + a 01 a 12 I pq2x2
I pq2y = a 21 a 02 I pq2y1 + a 01 a 22 I pq2y2
I pq2xj = Z xj exp[ i( V z1 h 1 + V z2 h 2 ) ] exp[ i( V 1 r 1 + V 2 r 2 ) ]d r 1 d r 2
I pq2yj = Z yj exp[ i( V z1 h 1 + V z2 h 2 ) ] exp[ i( V 1 r 1 + V 2 r 2 ) ]d r 1 d r 2
F( V z1 , V z2 ,ρ )= exp[ i( V z1 h 1 + V z2 h 2 ) ] =βexp( β 12 ρ )
Z xj exp[ i( V z1 h 1 + V z2 h 2 ) ] = ( 1 ) 3j i V z(3j) δ 2 ρ ξ cosαF( V z1 , V z2 ,ρ )
Z yj exp[ i( V z1 h 1 + V z2 h 2 ) ] = ( 1 ) 3j i V z(3j) δ 2 ρ ξ sinαF( V z1 , V z2 ,ρ )
I pq20 = a 01 a 02 πβ D 2 2 exp( D 2 | V c | 2 2 ) 0 ξdξexp( ξ 2 2 D 2 ) exp( β 12 ρ )Η( ξ )
I pq2xj = ( 1 ) 3j i δ 2 V z( 3j ) πβ D 2 2 exp( D 2 | V c | 2 2 ) 0 ξdξexp( ξ 2 2 D 2 ) ρ ξ exp( β 12 ρ )Λ( ξ )
I pq2yj = ( 1 ) 3j i δ 2 V z( 3j ) πβ D 2 2 exp( D 2 | V c | 2 2 ) 0 ξdξexp( ξ 2 2 D 2 ) ρ ξ exp( β 12 ρ )Φ( ξ )
Η( ξ )= 0 2π dαexp[ iξ| V d | 2 cos( αε ) ]
Λ( ξ )= 0 2π dαcosαexp[ iξ| V d | 2 cos( αε ) ]
Φ( ξ )= 0 2π dαsinαexp[ iξ| V d | 2 cos( αε ) ]
Η( ξ )=2π J 0 [ ξ 2 | V d | ]
Λ( ξ )=i2π J 1 ( ξ 2 | V d | )cosε
Φ( ξ )=i2π J 1 ( ξ 2 | V d | )sinε
I pq20 = π 2 D 2 a 01 a 02 βexp( D 2 | V c | 2 2 ){ exp( β 12 ) 0 ξ 0 ξdξexp( ξ 2 2 D 2 ) × J 0 [ ξ 2 | V d | ]+ ξ 0 ξdξexp( ξ 2 2 D 2 ) J 0 [ ξ 2 | V d | ] }
I pq2xj = ( 1 ) 4j V z( 3j ) π 2 D 2 δ 2 βcosεexp( D 2 | V c | 2 2 ){ exp( β 12 ) 0 ξ 0 ξdξ ×exp( ξ 2 2 D 2 ) ρ ξ J 1 [ ξ 2 | V d | ]+ ξ 0 ξdξexp( ξ 2 2 D 2 ) ρ ξ J 1 [ ξ 2 | V d | ] }
I pq2yj = ( 1 ) 4j V z( 3j ) π 2 D 2 δ 2 βsinεexp( D 2 | V c | 2 2 ){ exp( β 12 ) 0 ξ 0 ξdξ × exp( ξ 2 2 D 2 ) ρ ξ J 1 [ ξ 2 | V d | ]+ ξ 0 ξdξexp( ξ 2 2 D 2 ) ρ ξ J 1 [ ξ 2 | V d | ] }
ξ 0 = l c { ln( β 12 ln2ln[ 1+exp( β 12 ) ] ) } 1/2
J n ( z )= k=0 ( 1 ) k z 2k+n 2 2k+n k!( k+n )!
γ( a,x )= 0 x e t t a1 dt Γ( a,x )= x e t t a1 dt [Re a>0]
I pq20 = π 2 D 4 a 01 a 02 βexp( D 2 | V c | 2 2 ) k=0 ( 1 ) k ( | V d |D ) 2k 2 3k ( k! ) 2 ×[ exp( β 12 )γ( k+1, ξ 0 2 2 D 2 )+Γ( k+1, ξ 0 2 2 D 2 ) ]
I pq2s =[ ( a 11 a 02 V z2 a 01 a 12 V z1 ) V dx +( a 21 a 02 V z2 a 01 a 22 V z1 ) V dy ] π 2 D 6 δ 2 β ×exp( D 2 | V c | 2 2 ) k=0 ( 1 ) k | V d | 2k D 2k l c 2k+2 2 3k k!( k+1 )! ( l c 2 +2 D 2 ) k+2 [ exp( β 12 )γ( k+2, x 0 )+Γ( k+2, x 0 ) ]
E pqj s = ( a 0j + a 1j Z x + a 2j Z y )exp( i V zj h ) exp( i V j r )dS
exp[ i V zj h( r ) ] =exp( V zj 2 δ 2 /2 )=exp( β jj /2 )
Z x exp[ i V zj h( r ) ] = Z y exp[ i V zj h( r ) ] =0
E pqj s =π D 2 a 0j exp( β jj /2 )exp( D 2 | V j | 2 /4 )
Select as filters


Select Topics Cancel
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.