Introduction to Radar Imaging Margaret Cheney Outline Mathematical - - PowerPoint PPT Presentation
Introduction to Radar Imaging Margaret Cheney Outline Mathematical - - PowerPoint PPT Presentation
Introduction to Radar Imaging Margaret Cheney Outline Mathematical model Image formation time domain viewpoint frequency domain viewpoint (for small scenes) Approximating targets by point clouds SAR interferometry
Outline
- Mathematical model
- Image formation
- time domain viewpoint
- frequency domain viewpoint (for small scenes)
- Approximating targets by point clouds
- SAR interferometry
Maxwell’s equations —-> scalar wave equation Green’s function + Born approximation
p(t, xr; xs) ∝ Z ρ(y)f 00[t − τ(y, xs) − τ(y, xr)] |y − xs||y − xr| dy
standard (monostatic) SAR:
Mathematical Model p(t, s) ∝ Z ρ(y)f 00 (t − 2Rs,y/c0) R2
s,y
dy
τ(y, x) = |y − x| c0 = |γ(s) − y| c0
y Rs,y xr = xs = γ(s)
figures from Brett Borden, Naval Postgraduate School
Image formation
standard (monostatic) SAR array imaging
I(y) = X
xr,xs
data(t = τ(y, xs) + τ(y, xr), xr; xs)
why does this work?
= X
s
data
- t = 2Rs,y/c0, γ(s)
Imaging from a single viewing position
Imaging from a single view Example with 3 scatterers
Imaging from two views
Imaging from three views synthetic aperture
Frequency domain viewpoint
p(t, s) ∝ Z ρ(y)f 00 (t − 2Rs,y/c0) R2
s,y
dy
time-domain model
P(ω, s) ∝ Z ρ(y)ω2F(ω)e−2iωRs,y/c0 R2
s,y
dy
Z . . . e−iωtdt
Fourier transform in t far-field approximation
|γ| |y|
Rs,y = |γ(s) − y| ≈ |γ| − b γ · y + · · ·
k = ω c0
to form image, invert Fourier transform!
b γ = γ |γ|
P(ω, s) ∝ Z ρ(y)e−2ikb
γ(s)·ydy
Approximating targets by point clouds P(ω, s) ∝ Z ρ(y)e−2ikb
γ·ydy
k large -> use geometrical optics main contributions are from corners, edges, and specular points
Interferometry p(t, s) ∝ Z ρ(y)f 00 (t − 2Rs,y/c0) R2
s,y
dy p(t, s) ∝ Z ρ(y)eiω0(t−2Rs,y/c0) a(t − 2Rs,y/c0) R2
s,y
dy ρ(y) = ˜ ρ(y1, y2)δ(y3 − h(y1, y2)) f(t) = a(t)eiω0t
narrowband slowly varying (complex) amplitude
a =
scattering takes place on surface
yT = (y1, y2, 0) Rs,y = |yT + hˆ e3 − γ| = |yT − γ| | {z }
Rs,yT
+ h(yT )ˆ e3 · \ yT − γ | {z }
d(yT )
y = yT + h(yT )ˆ e3 p(t, s) ≈ Z h ˜ ρ(yT )e2ik0d(yT )i eiω0(t−Rs,yT /c0) a(t − Rs,yT /c0) R2
s,yT
dyT
target phase encodes height information!