🛰 SAR · Satellite Imagery
Foundations — the physics of signal & image

Synthetic aperture principle

matched filterAzimuth compressionwiki/sar-synthetic-aperture
TL;DR

Synthetic Aperture is a technique that moves a small antenna along the satellite track, observes the same target hundreds to thousands of times, and sums those complex signals after phase alignment — producing the same effect as a kilometers-wide giant antenna. As the satellite moves, the range to the target (R1, R2, …) changes, and that range difference shows up as a phase difference; the process where a processor that knows it is the same target corrects the phase and adds them up is exactly Azimuth Compression (a matched filter). Because the signals must be summed with not just amplitude but phase aligned, complex (I/Q) data is essential, and since this step is already completed inside SNAP's RAW→SLC stage, SLC, Interferogram, and Coherence are all products of synthetic aperture.

Why it is needed — the antenna dilemma

  • Azimuth resolution improves the longer the antenna is, but you cannot mount a 10km antenna on a satellite (Sentinel-1's real antenna is 12.3m).
  • So instead of physically enlarging the antenna, the core idea of synthetic aperture is to combine observations gathered while the satellite moves, creating a virtual ultra-large antenna.
  • It synthesizes — via the satellite's trajectory rather than a physical antenna — the same principle by which a larger lens (aperture) yields higher resolution in an optical camera.

Mechanism — range difference becomes phase difference

  • As an active sensor the satellite repeats 'beep! → echo → received' over and over while continuously moving along its track.
  • A single target is observed hundreds to thousands of times as the satellite passes, the range to it differs slightly at each position, and a different range means a different phase in the returning signal.
  • The key is that the processor knows this phase variation comes from the same target, so when each observation is summed after phase correction (a matched filter), the signal concentrates strongly at a single point.
📡Real antenna movesSatellite travels along track t1·t2·…·tN
🛰️Hundreds–thousands of observationsSame target, ranges R1≠R2≠…≠RN → phase differs per position
🔄Phase correction and summationAzimuth Compression / matched filter
🎯Virtual giant antennaEnergy concentrated at one point → high azimuth resolution → SLC
Synthetic aperture processing flow: from antenna motion to a virtual giant antenna

Small antenna vs synthetic aperture, and why complex numbers

  • With only the small real antenna, the beam spreads wide like a flashlight, blurring which target is which, and azimuth resolution can degrade to hundreds of meters.
  • After synthetic aperture, the beam focuses like a laser pointer so the signal is strong only at a specific position, yielding an azimuth resolution of about 20m.
  • Synthetic aperture is not a plain sum but an addition with phase aligned; for '1+1+1+1=4' to hold the directions (phases) must match, which is why complex (I/Q) data is essential.
  • On the complex plane phase is an angle (A·e^(jθ)), and phase correction is a rotation — a single multiplication — which makes the summation easy.
Beam shapeResult
Small real antenna onlySpreads wide like a flashlightBlurry target — azimuth worsens to hundreds of m
After synthetic apertureFocused like a laser pointerStrong only at one position — azimuth ~20m
Small real antenna only vs after synthetic aperture

Sentinel-1 numbers and where it sits in SNAP

  • Sentinel-1 operates at 5.405 GHz with a wavelength λ of about 5.6 cm, a real antenna length of about 12.3 m, and an IW swath of about 250 km.
  • With the real antenna alone azimuth resolution is on the order of hundreds of meters, but synthetic aperture brings the final value down to about 20 m (including TOPS and operational-mode effects).
  • In SNAP processing, synthetic aperture corresponds to the Azimuth Compression step in RAW → Range Compression → Azimuth Compression (synthetic aperture) → SLC.
  • SLC, Interferogram, Coherence, DSM, DInSAR, and SBAS all use the completed synthetic-aperture result, so if synthetic aperture fails the SLC itself is not produced properly.
Pitfalls & gotchas

In practice you rarely touch synthetic aperture directly, but it is the starting point when explaining why azimuth resolution is worse than range. Forgetting that SLC, Interferogram, and Coherence are all synthetic-aperture products makes it easy to miss the causal link that a synthetic-aperture failure means an SLC-generation failure.

The topic cards on this page are compiled from the Brain Trinity wiki. The original wiki can be demoed live in an interview.Back to study log