1. Programme for Research-Development-Innovation for
Space Technology and Advanced Research - STAR
Compressive THz Imaging and Hadamard Spectroscopy for Space Applications THz Imaging
Florin Garoi
Romanian Space Week , May 2014, Bucharest, Romania

2. Partners Coordinating organization Partner organization
National Institute for Laser, Plasma and Radiation Physics (INFLPR)
University Politehnica of Bucharest – Research Center for Spatial Information (CEO Space Tech)
409 Atomistilor Street, PO Box MG-36, Magurele, Ilfov, Romania
1-3 Iuliu Maniu Bd, Bucharest, Romania
Project manager
Partner team leader
Florin Garoi
Daniela Coltuc
INFLPR
CEO Space Tech
Tel/Fax:
Tel:
Web:
Web:

3. The project Short description of the project Project goal
spectroscopy and imaging at THz wavelength range (0.1 – 30 THz);
Hadamard spectroscopy
Compressive sensing (CS) imaging
Project goal
Experimental model of Hadamard spectrometer and CS imaging system in THz domain, for Space Applications

4. * with gray are already completed activities
Work plan of the project
* with gray are already completed activities
Phase
Activities
Partner
Deadline I
State of the art in THz imaging and Hadamard spectroscopy
State of the art in compressive sensing (CS) and Hadamard spectrocospy
Compressive sensing and Hadamard spectroscopy basics
Financial and Quality Management
INFLPR
UPB
Dec 15, 2012 II
Concept of THz image spectrometer and simulation of imaging and spectroscopy
THz image spectrometer as and-to-and system
Modeling and simulation of imaging and spectroscopy functions
Reporting and dissemination
June 30, 2013
III
Experimental model of THz image spectrometer
Design and implementation of the THz data acquisition and DMD
Characterization of the components in the experimental setup
Integration of the parts and testing of electrical components of the experimental setup
Statistical analysis of the measured THz signals and elaboration of appropriate algorithms for CS
Dec 15, 2013 IV
Functions development for THz image spectrometer
Definition of the CS and spectroscopy specifications
Design and implementation of the CS and spectroscopy
Integration of CS and spectroscopy in the experimental model
Dec 15, 2014 V
THz image spectrometer validation and evaluation
Laboratory simulator for ESA applications of interest in THz domain
Study regarding development of technical perspectives
Nov 18, 2015

5. Implementation status
Characterization of the components in the experimental setup
Experimental model v1.0:
FIR 100 laser (Edinburgh Instruments); l = mm, mm, and 163 mm
Mirrors
Dispersive component:
reflective blaze diffraction grating
transmission diffraction grating (wire or machine cut)
prism
Digital Micromirror Device (DMD)
Detector
Software:
LabView CS
Laser
CO2 section: 80 lines between 9.1μm and 10.9μm
50W on the strongest lines
CO2 laser output is coupled into the FIR laser via two steering mirrors and a ZnSe focusing lens
FIR section:
118.83mm (150mW), 133.1mm (1mW), and 163mm (36mW)
Installation and training

6. Implementation status
Optical components
Off-axis parabolic mirrors:
f = 250mm (Laser Beam Products, UK)
Lenses:
f = 50mm, 100mm, 200mm (Tydex, Russia)
Reflective blaze diffraction grating:
Brass (CNC EUROMOD-P), coated with 50nm layer of gold (Varian RF magnetron sputtering)
Dimensions: 20mm × 20mm × 5mm
Grating pitch: 120 mm  not feasible
300 mm  in the making
Cross-section of the blaze grating
Grating cross-section
Blaze wavelength as a function of the incidence angle.

7. Implementation status
Wire transmission diffraction grating:
Brass frame and nickel wire
Dimensions: 50mm × 50mm × 10mm
Grating active area: 40mm × 40mm
Wire diameter: 200mm
Grating pitch: 400mm
Polyethylene prisms:
Prism angle: 60, 50 and 40
Wavelength as a function of
diffraction angle (first diffraction order)
Measured with Tera View systems, TPS Spectra 3000
Deviation d versus angle of incidence a

8. Implementation status
Digital Micromirror Device, DAQ and Detector:
Projection of masks and acquisition  integrated in LabView
Tested in visible range of the spectrum
Testing with various random matrices generation for CS
Find a sensing matrix through random methods that accurately reproduces a given image and scaling for a finer resolution with more samples or larger sensors coverage; for example Low Density Parity Check (LDPC) matrices
Project start
TRL 1 – Basic principles observed and reported
Lowest level of technology readiness. Scientific research begins to be translated into applied research and development
Present
TRL 2 – Technology concept and/or application formulated
Once basic principles are observed, practical applications can be invented and R&D started. Applications are speculative and may be unproven
TRL 3 – Analytical and experimental critical function and/or characteristic proof-of-concept
Active research and development is initiated, including analytical/laboratory studies to validate predictions regarding the technology
Intended to be achieved
TRL 4 – Component and/or breadboard validation in laboratory environment
Basic technological components are integrated to establish that they will work together

9. Results Physical Fourier encoding of optical data
Preliminary tests of the dispersive elements, in visible and THz range
Experimental setup
Initial objects
Decoded images

10. Results
One Pixel Camera for THz Image Acquisition Design and Tests of Camera Software
EXPERIMENTAL RESULTS
ACQUISITION PRINCIPLE: Compressive Sensing (CS)
SCOPE: finding the appropriate couple
Sensing Matrix – Transform for CS Algorithm
METHOD: couple evaluation by Rate-Distortion curve
EXPERIMENTS:
Test images (numerical):
Cameraman and Lena
Tested Sensing matrices: Binary Random
Binary Sparse
LDPC (Low Density Parity Code)
Tested Transforms: Discrete Cosine Transform
Wavelet Transform
TV (Total Variation)
Rate-Distortion curves for TV
SOFTWARE: CS Toolbox at
L1 Magic at
FOUND SOLUTION: LDPC with TV
Rate-Distortion curves for LDPC

11. Project contribution to STAR Programme
Project’s contribution to the goal of the STAR Programme
(how the project contributes to the increasing of the capacity for organizations involved to participate in ESA Programmes)
The project addresses Basic Technology Research Programme (TRP), one of ESA’s activities regarding technology and research → may facilitate strong and long-term relations between Romanian entities and ESA.
Context and contribution to ESA Programmes
(please specify how the project activities can contributes to present and future ESA programmes)
We hope that our research on THz imaging with CS and Hadamard spectroscopy contribute to the development of new instruments required for the current and long-term ESA science directorate programme.

12. Dissemination Publications and Conferences
Rate-Distortion Performance of Compressive Sensing in One Pixel Camera, Mihai Petrovici, Daniela Coltuc, Mihai Datcu, and Vasile Tiberius, Submitted to IEEE Signal processing Letters 
Physical Fourier encoding and compacting of optical data, Petre Catalin Logofatu, Florin Garoi, Victor Damian and Cristian Udrea, Submitted to Optical Engineering
Terahertz range complex refractive index determinations for liquids using ATR, Adrian Dobroiu and Petre Catalin Logofatu, Submitted to Optical Engineering
Introduction to Compressive Sampling and applications in THz Imaging, Daniela Coltuc, Invited paper to ATOM-N 2014 Conference (Constanta, August 21-24, 2014)
Dispersive elements for THz domain, V. Damian, Mihaela Bojan, G. D. Chioibasu, F. Garoi, P. C. Logofatu, L. Mihai, C. Udrea, I. Urzica, T. Vasile, and C. Viespe, will be presented at INDLAS Conference (Bran , May 19-23, 2014)
Synthetic aperture and scarsity analysis methods for THz imaging, Mihai-Alexandru Petrovici, will be presented at the European Conference on Computer Vision (Zurich, September 6-12, 2014)
L1 minimization applied to the simple case of a sparse signal consisting in a sum of sinusoids, Petre Catalin Logofatu, will be presented at ATOM-N Conference (Constanta, August 21-24, 2014)
Technical Notes within the consortium
(available upon request)
For up-to-date info, please visit our webpage:

13. Conclusions Conclusions
All tasks and activities completed as scheduled
Most important results:
Installation and training for the THz laser
Simulation of the experimental setup v1.0
Characterization of some of the components
Definition of the CS matrices and algorithms for data processing