UNCLASSIFIED CLASSIFICATION The NRL Multi Aperture SAR (NRL MSAR): System Description and Recent Results Luke Rosenberg Defence Science and Technology Organisation, Australia Mark Sletten, Naval Research Laboratory, USA 1
UNCLASSIFIED Outline • Motion in SAR imagery • Single vs. Multi-aperture SAR • The Velocity SAR algorithm for focussing moving scatterers • Demonstration of the VSAR algorithm using the NRL FOPAIR • Initial results from the Airborne MSAR system • Enhanced VSAR • Future plans Acknowledgements: • Naval Research Laboratory, Remote Sensing Division: Mark Sletten, Steve Menk, Jakov Toporkov, Bob Jansen • Naval Research Laboratory, Radar Division: Raghu Raj, Denny Baden 2
UNCLASSIFIED Effects of Scene/Target Motion on SAR Signatures • Relative motion between platform and scene is fundamental to SAR • SAR processors assume scene is stationary: Scene motion causes distortion • Constant radial motion: azimuthal offsets, a.k.a. “train off the track” distortion • Radial acceleration and azimuthal motion: azimuth defocusing • Issue is significant for marine applications, since complex motion is pervasive • Signatures not only displaced, but smeared as well Real Aperture Radar Image SAR Image (emulated) NRL FOPAIR Imagery, Small boat on the Chesapeake Bay 3
UNCLASSIFIED Formation of a Standard SAR Image Image Cars Road Azimuth 1 phase center Time Space Synthetic aperture 4
UNCLASSIFIED Formation of an MSAR Image Stack Image Cars Road Azimuth M phase centers Time Space 5
UNCLASSIFIED Formation of an MSAR Image Stack Image Time Stack 𝑢 = 𝑢 0 Cars Road 𝑢 = 𝑢 0 + ∆ 𝑢 M phase centers 𝑢 = 𝑢 0 + 2∆ 𝑢 𝑢 = 𝑢 0 + 3∆ 𝑢 Time 𝑢 = 𝑢 0 + (𝑁 − 1)∆ 𝑢 Space Azimuth • Images look the same: motion information is in the phase of the complex pixels • Images look the same: motion information is in the phase of the complex pixels 6
UNCLASSIFIED Velocity SAR (VSAR) Processing • Doppler processing converts the image time-stack into a velocity stack 𝑆 𝑞 𝑤 𝐸𝑝𝑞 corrects azimuthal misplacement • Shifting each velocity image by 𝑊 • An incoherent sum down the corrected velocity stack forms a single corrected image Time Stack Velocity Stack Shifted Velocity Stack Doppler Frequency/velocity FFT Azimuth Azimuth Corrected Image (Incoherent sum) Azimuth Azimuth 7
UNCLASSIFIED NRL FOPAIR • NRL Focused Phased Array Imaging Radar (NRL FOPAIR) • Updated version of UMass FOPAIR ( McIntosh and Frasier , 1995) • Mimics a SAR: Receive array elements rapidly and sequentially scanned • Generates image time- stacks with a high frame rate (780 fps “movies”) • X-band (9.85 GHz) fully polarimetric, 200 MHz BW (0.75 m resolution) • 16-module receive array easy to reconfigure 8
UNCLASSIFIED FOPAIR as an MSAR Test Bed MSAR FOPAIR d eff M Apertures … Time Time Space Space 9
UNCLASSIFIED First Reported Demonstration of VSAR-Based Signature Correction • NRL FOPAIR imagery of a small boat used to demonstrate VSAR signature correction Sletten , IEEE Trans. Geoscience Remote Sens ., Vol. 51, No. 5, May 2013 SAR Image (emulated) VSAR Image (emulated) 10
UNCLASSIFIED NRL MSAR Basic Specifications • X-band (9.875 GHz CF) • Bandwidth: 220 MHz • Waveform: LFM, both up and down chirps • Peak and average power: 1.4 kW, 210 W • Phase centers: 32 along-track • Polarization: VV • Platform: Saab 340 • IMU: Novatel • Data recorder: NRL custom-built, 4-channel, 800 MB/s sustained ~ 1200 m 45 ° 22 ° ~ 1.5 km 11
UNCLASSIFIED NRL MSAR Aircraft and Radome Saab 340 Tx Down-chirp Tx Up-chirp Rx 1-16 Novatel IMU (behind Rx modules) 12
UNCLASSIFIED 32 Phase Center Array 2 Transmit Horns d=10.5 cm 16 Physical Receive Elements d eff 32 Resulting Phase Centers d eff ≈ d/2 = 5.25 cm • Use two transmit antennas to double number of phase centers • Minimum and maximum unambiguous velocities, assuming VSAR-type processing: V V p p v 10 m / s v 0 . 7 m / s At V p =70 m/s (Saab 340) max 4 d 2 d M min eff eff • Cycle through all 32 combinations of Tx and Rx antennas in 320 microsec (8 pulses) 13
UNCLASSIFIED Antenna Switching Schematic Up chirp transmit antenna Down chirp transmit antenna D U 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Receive Elements Switches Data Data Data Data acquisition acquisition acquisition acquisition channel 1 channel 2 channel 3 channel 4 14
UNCLASSIFIED Pulse 1 Up-chirp Receive elements 1, 9, 17, 25 U 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Receive Elements Switches Data Data Data Data acquisition acquisition acquisition acquisition channel 1 channel 2 channel 3 channel 4 15
UNCLASSIFIED Pulse 2 Down-chirp Receive elements 1, 9, 17, 25 D 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Receive Elements Switches Data Data Data Data acquisition acquisition acquisition acquisition channel 1 channel 2 channel 3 channel 4 16
UNCLASSIFIED Pulse 3 Up-chirp Receive elements 3, 11, 19, 27 U 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Receive Elements Switches Data Data Data Data acquisition acquisition acquisition acquisition channel 1 channel 2 channel 3 channel 4 17
UNCLASSIFIED Pulse 4 Down-chirp Receive elements 3, 11, 19, 27 D 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Receive Elements Switches Data Data Data Data acquisition acquisition acquisition acquisition channel 1 channel 2 channel 3 channel 4 18
UNCLASSIFIED Pulse 5 Up-chirp Receive elements 5, 13, 21, 29 U 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Receive Elements Switches Data Data Data Data acquisition acquisition acquisition acquisition channel 1 channel 2 channel 3 channel 4 19
UNCLASSIFIED Pulse 6 Down-chirp Receive elements 5, 13, 21, 29 D 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Receive Elements Switches Data Data Data Data acquisition acquisition acquisition acquisition channel 1 channel 2 channel 3 channel 4 20
UNCLASSIFIED Pulse 7 Up-chirp Receive elements 7, 15, 23, 31 U 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Receive Elements Switches Data Data Data Data acquisition acquisition acquisition acquisition channel 1 channel 2 channel 3 channel 4 21
UNCLASSIFIED Pulse 8 Down-chirp Receive elements 7, 15, 23, 31 D 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Receive Elements Switches Data Data Data Data acquisition acquisition acquisition acquisition channel 1 channel 2 channel 3 channel 4 22
UNCLASSIFIED Inaugural NRL MSAR Deployment • 30 flight hours over the span of 2+ weeks, September 2014. • Based out of Newport News VA. • After a difficult installation, system worked remarkably well. Some issues: • Due to placement of transmit horns, only 28 unique phase centres. • Mismatch with up-chirp / down-chirp waveforms - produced low image coherence. Current VSAR results are restricted to 16 phase centres. • Two subjects of study • Oregon Inlet on the Outer Banks of NC • Imaged boats of opportunity, waves, currents, vehicles. • Used linear flight patterns (i.e. strip-map). • Cooperative vessels in the Southern Chesapeake Bay. • Imaged 30 different vessels, both stationary and moving (0-50 kts). • Used both linear and circular flight patterns. 23
UNCLASSIFIED Inaugural NRL MSAR Deployment Oregon Inlet, NC Outer Banks VSAR analysis region 24
UNCLASSIFIED First VSAR Analysis Shoaling Animation: waves Click to start Northbound Southbound Vehicles • VSAR processing significantly reduces smearing of shoaling waves • (Faint) vehicle signatures shifted back to bridge 25
UNCLASSIFIED Velocity Image Movie Each image shifted by 𝑺 Animation: 𝑾 𝒒 𝒘 𝑬𝒑𝒒 to Click to start correct azimuthal displacement Northbound Southbound Vehicles • Vehicle signatures much more visible than in previous composite image, due to Doppler filtering inherent in VSAR processing • Vehicle speeds projected onto bridge are 64 and 48 mph (speed limit 55 mph) 26
Recommend
More recommend
