Talk:Synthetic aperture radar

From Wikipedia, the free encyclopedia

Contents 1 Maximum Resolution 2 Doppler Beam Sharpening 3 Doppler Beam Sharpening 4 No. 5 ISAR 6 The real answer 7 UWB SAR

[edit] Maximum Resolution

First of all I found it amazing that holograms were being used to perform analogue computations on datasets. How long ago was this being done? Does anyone know?

Seondly the resolution quoted in the article is 0.6m. If we were doing it at 0.6m resolution back then what's the maximum resolution attainable now? —Preceding unsigned comment added by 24.80.86.98 (talk) 18:58, 22 January 2008 (UTC)

[edit] Doppler Beam Sharpening

Is this really compatible with SAR? The Doppler shift from targets ahead/behind seems like it is actually the same effect as the phase shift that is used to assemble multiple samples. For a sidelooking radar that is not a SAR, it's clear this could help. --Andrew 18:54, Dec 14, 2004 (UTC)

[edit] Doppler Beam Sharpening

Hmmm... I'm not sure if I understand your question, but let me say (from many years of experience) that most SAR radars use doppler beam sharpening.

Let me first say that it is EXTREMELY DIFFICULT to separate frequency shift from phase shift... at the time the shift is occurring. After the fact it's not a problem, but with a dynamic signal, it's nigh impossible to separate the two... This "reality" is often used with FM radios - in that the same receive chain detects either PM (phase modulation) or FM with little difference...

Now as to the whole issue of dopper shifts... maybe it helps to separate the different contributing components (simplified). First, there are the two large scale doppler components: motion of the imaging platform and motion of "point" targets in the area being imaged. Normally (when MTI is not the intended capability) you'd like the point target to have no motion. Then there are the two "microdoppler" components: vibrations that cause spreading of the doppler signature - either by the imaging platform or by any point target vibrations. Unless you are taking advantage of microdoppler signature analysis then you'd just as soon these components were zero. Finally there are the periodic rotational/translational components - think of these as periodic movements on the target or imaging platform... actually used with benefit in ISAR imaging of ships (waves and swells cause regular rotational and translational displacements).

Now, if the goal is GROUND/SURFACE imaging... then typically... all the doppler components except the imaging platform - are assummed to be zero (as I said, simplified... we ignored yaw, roll, and pitch changes by the imaging platform - which all make life more difficult)... This has proven to be a resonable assumption in real world SAR systems. Never-the-less, imaging platforms do record/transfer the three space coordinates, three space accellerations, three space pitch/roll/yaw rates, and three space pitch/roll/yaw accelerations. GA maps also play a roll (GA = gravitational anomaly) to ensure the Z accelerations are normalized...

A complicated but interesting subject in itself. Anyway, many SAR systems treat the phase change as a frequency change by using DFT/FFT processing to determine the doppler shift amounts. This means that given a single range gate time period, the number of actual bins relates to the processing power available and the phase/frequency resolution (A/D speed, bits of resolution, linearity)... Obviously there are diminishing returns while working with all this that are fed by real energy beam width... i.e.... if the transmitted signal is way down when 15 degrees off orthonormal... then why bother to process data that would have doppler shifts that correlate to that far off axis... think in terms of basic trig... if the range gate being processed is 100 km away (60+ miles), the system has a ten meter "resolution" - - and the real beamwidth is 20 degrees... (+/- 10 degrees)... then the total number of bins that needs to be processed (tan "theta" = Opposite/Adjacent... since we know theta = 20 degrees and the adjacent side is 100,000 meters (100 km)... then the Opposite side is approximately 35,000 meters long... (simple triangle)... with 10 meter bins... then we're looking at 3,500 angular bins... just along that one range resolution cell... that must map into 3,500 discrete phase/frequency shifts... or 1750 discrete upshifted bins and 1750 discrete downshifted bins...

Anyway... as I digressed WAY TOO MUCH... Doppler beam sharpening... is exactly how many SAR systems actually do their imaging... (mapping mode - - - spot modes are another story)....

[edit] No.

Doppler beam sharpening is also called "unfocussed SAR" and it means that everything with a doppler shift >x is filtered out - so everything that is out of the center of the main beam is discarded. It is a simplified version of SAR.

[edit] ISAR

What's an Inverse Synthetic Aperture Radar? [1] Ojw 10:56, 6 October 2005 (UTC)

Inverse SAR uses the motion of the target to create a image. We plot range and doppler shift. Ships at sea can be imaged very easily, but since the velocity is not known ahead the images are pretty distorted. Dominick 11:40, 6 October 2005 (UTC) reference for you (to be added later) Dominick 11:47, 6 October 2005 (UTC)

[edit] The real answer

I want to point-out that my answer is more valid not because it has more merit or a greater factual basis but because it mine..... Just kidding, However these are my thoughts.

The processing is similar but the techniques are different. In fact they are opposites of each other; DBS integrates within a beam and SAR integrates out with the beam (or near enough). So what this means is that with DBS is synthesises a smaller beam width by taking pulse intervals within the beam spread and SAR takes pulse sample with a small overlap synthesising a larger antenna. The net result is the same and this is why the processing is similar for strip map and spot light but the techniques are opposite to each other. The difference is notable when you consider Scanned DBMS, or when you calculate the PRF and samples requires against a resolution for a given scenario.

[edit] UWB SAR

The section on UWB SAR contains two false statements that I will correct:

First, "Interpretation of the [narrow band radar] results is also eased by the fact that the material response must be known only in a narrow range of frequencies", is false in that the frequency response information is still present in a UWB echo, and can easily be considered in processing.

Strangely, the writer goes on to admit this fact, and presents it as a disadvantage!...

Second, "but much of the information is concentrated in relatively low frequencies (with long wavelengths). Thus such systems require very large receiving apertures to obtain correspondingly high resolution along the track." is also false. The large apertures are "required" only if one desires to detect that low-frequency information--it certainly isn't "necessary" to retrieve that information just because UWB SAR happens to provide the information while narrow-band SAR does not! Fleem (talk) 11:56, 9 May 2008 (UTC)