Navigation auf uzh.ch

Suche

Department of Geography Remote Sensing of Water Systems

Synthetic Aperture Radar (SAR) imagery uses radar signals to generate high-resolution images of Earth's surface, regardless of weather or lighting conditions. By measuring the time and intensity of radar pulses reflected from the ground, SAR captures detailed information on terrain, structures, and changes over time, making it essential for applications like environmental monitoring, disaster response, and surveillance.

The SARLab (Computational SAR) at RSL focuses on the understanding of the entire data processing chain of airborne and drone-based synthetic aperture radar systems: from SAR campaign planning through raw data focusing, calibration and validation to value-adding products.

Mission statement: Gaining knowledge through high quality SAR images and products.

 

Multifrequency Polarimetric SAR

Polarimetric SAR (PolSAR) is a radar imaging technique that utilizes various polarization combinations (e.g. HH, HV, VH, VV) to examine how objects scatter electromagnetic waves. By configuring the transmitting and receiving antennas in multiple polarization modes, PolSAR can distinguish between different scattering mechanisms—such as even-bounce (from flat surfaces), odd-bounce (from vertical structures like buildings), and volume scattering (from vegetation).

 

QuadPolSAR:

  • Enhanced Object Information: QuadPol helps in distinguishing between diverse land features, vegetation types, and man-made structures.

 

Multifrequency SAR utilizes multiple radar frequency bands to capture images of the same area, each band interacting uniquely with surface features. X-band provides high-resolution imagery but has limited penetration, making it well-suited for surface-level detail. C-band offers moderate penetration, balancing resolution and depth for versatile observations. L-band penetrates more deeply, ideal for exploring subsurface features and imaging through dense vegetation. By combining these frequencies, multifrequency SAR delivers a multi-layered view, enhancing our understanding of both surface and subsurface characteristics across diverse landscapes.

 

Multifrequency SAR:

  • Depth and Subsurface Analysis: different frequencies penetrate various depths. For example, X-band (short wavelength) primarily captures surface details, C-band penetrates vegetation moderately, and L-band (long wavelength) penetrates more deeply into the vegetation canopy and soil.

Here are some examples of images acquired with DLR's F-SAR sensor in 2019, Memmingen (Germany).

  • Pauli RGB Composite at X-band (R=HH-VV, G=HV+VH, B=HH+VV)

    This composite illustrates how various polarization combinations enhance distinct features of objects. The red channel (R) represents even-bounce scattering, which is commonly associated with walls and vertical structures. The green channel (G) indicates volume scattering, typically observed in vegetation or complex structures. The blue channel (B) shows odd-bounce scattering, which occurs in corner reflectors and flat surfaces.

  • Pauli RGB Composite at C-band (R=HH-VV, G=HV+VH, B=HH+VV)

    This composite Highlights how different polarization combinations enhance different object features

  • Pauli RGB Composite at L-band (R=HH-VV, G=HV+VH, B=HH+VV)

    This composite highlights how different polarization combinations enhance different object features.

  • Multifrequency RGB Composite with the HH images (R=X-band HH, G=C-band HH, B=L-band HH)

    This composite demonstrates how different frequencies highlight varying target features.

  • Multifrequency Odd-Bounce Composite (R=HH+VV at X-band, G= HH+VV at C-band, B= HH+VV at L-band)

    This composite highlights buildings or vertical structures across frequencies.

  • Multifrequency Volume Scattering Composite (R=HV+VH at X-band, G= HV+VH at C-band, B= HV+VH at L-band)

    This composite shows vegetative canopy variations.

  • Multifrequency Even-Bounce Composite (R=HH-VV at X-band, G= HH-VV at C-band, B= HH-VV at L-band)

    This composite differentiates flat surface areas, like water bodies and ground planes.

Ground Moving Target Indication (GMTI)

SAR-GMTI (Synthetic Aperture RADAR - Ground Moving Target Indication) is a RADAR technique used to detect moving targets within SAR images, which are typically designed to capture high-resolution ground imagery. The signature of a moving target (e.g., cars, trains, vessels) is usually blurred and displaced in position due to the relative motion between the SAR sensor and the target, which is not accounted for during traditional SAR focusing. Moving target refocusing techniques are designed to compensate for this relative motion, thereby "refocusing" the image to correct the blur and provide a clearer, high-resolution representation of the moving object.

Furthermore, when multiple receiving channels are available, it is possible to further process the refocused images, merge their information, and derive a three-dimensional point cloud of the target under analysis.

Vehicle Detection using AI

One application of SAR imagery is vehicle detection, which can significantly enhance surveillance capabilities. To achieve this, we trained a deep learning model to detect cars. Here are some of the results:

Group members

We welcome you to reach out to us anytime—don't hesitate to get in touch!

Open Positions and Theses