Remote Sensing and Aerospace Technologies

Transcription

1 Remote Sensing and Aerospace Technologies Stefano Baronti Institute of Applied Physics Nello Carrara CNR Department of Engineering, ICT and Technologies for Energy and Transport Annual conference, CNR Auditorium Pisa (Italy), 16 th 18 th November 2015

2 Design Principles

3 Sensor Panchromatic (PAN) 1 Band

4 Sensor Multispectral (MS) Few bands (3-4)

5 Sensor Superspectral Tens of bands

6 Sensor Hyperspectral Hundreds of bands

7 Sensor Ultraspectral Thousands of bands

8 Whisk-broom Scanners Along track direction Dispersion element Across track direction Linear array detector

9 Push-broom Scanners Spectral direction - Along track direction - y Across track direction - x

10 Time Delay & Integration (TDI)

11 t0

12 t1

13 t2

14 t3

15 TDI Advantages Increased signal-to-noise ratio Photo-response uniformity (averaging/destriping) Enhanced photo-sensitivity without sacrificing output data rate Dark current uniformity (averaging) Electronic exposure control

16 Compressive Sensing The key idea of compressive sensing (CS) is to directly acquire compressed data. acquire measurements, not samples. We

17 Compressive Sensing The key idea of compressive sensing (CS) is to directly acquire compressed data. acquire measurements, not samples. We Shannon-Nyquist Bandlimited signals can be sampled and reconstructed. N samples are obtained. The compression retains only K best coefficients (K<<N) typically expressed in a transformed domain.

18 Compressive Sensing The key idea of compressive sensing (CS) is to directly acquire compressed data. acquire measurements, not samples. We Shannon-Nyquist Compressive Sensing Bandlimited signals can be sampled and reconstructed. N samples are obtained. The compression retains only K best coefficients (K<<N) typically expressed in a transformed domain. Sparse signals can be undersampled and recovered. Compressive sampling acquires K<<N measurements and no further compression is needed.

19 Compressive Sensing: Single-Pixel Camera Detail of the baboon test tmage (256 X 256 pixels) and CS reconstruction image using 6600 measurements (

20 INSTRUMENTS

21 PRISMA PRecursore IperSpettrale della Missione Applicativa

22 PRISMA Characteristics Panchromatic camera Hyperspectral camera Type n.a. Push-broom Number of Channel Head configuration n.a. 2 imaging spectrometer with prisms Field of view and swath 2, km 2, km Across-track pixels: Ground Sample Distance: km km Spectral interval: nm nm Spectral resolution n.a. 12 nm SNR: > nm: > nm: > 100 MTF: > Nyquist frequency 0.3 across track, 0.17 along Nyquist frequency

23 PRISMA Characteristics Panchromatic camera Hyperspectral camera Type n.a. Push-broom Number of Channel Head configuration n.a. 2 imaging spectrometer with prisms Field of view and swath 2, km 2, km Across-track pixels: Ground Sample Distance: km km Spectral interval: nm nm Spectral resolution n.a. 12 nm SNR: > nm: > nm: > 100 MTF: > Nyquist frequency 0.3 across track, 0.17 along Nyquist frequency

24 OPTIMA Project: Data Fusion (Pansharpening) Pansharpening is the process of extracting spatial details from a high spatial resolution image (Panchromatic) and adding these details to low resolution multispectral/hyperspectral bands, using a detail injection model. The result is a multispectral image with the same characteristics but with high spatial resolution.

25 OPTIMA Project: Data Fusion (Pansharpening) Pansharpening is the process of extracting spatial details from a high spatial resolution image (Panchromatic) and adding these details to low resolution multispectral/hyperspectral bands, using a detail injection model. The result is a multispectral image with the same characteristics but with high spatial resolution. One of the goals of the project has been to establish the relationship between the spectral range of the panchromatic and the properties of the pansharpened multispectral image.

26 Two methods are tested: Component Substitution (CS) Multiresolution Analysis (MRA)

27 Relative Response Wide Panchromatic Wavelength (nm)

28 Relative Response Narrow Panchromatic Wavelength (nm)

29 Original Wide MRA Narrow MRA Exp Wide Pan Narrow Pan

30 Original Wide MRA Narrow MRA Exp Wide Pan Narrow Pan

31 Original Wide MRA Narrow MRA Exp Wide Pan Narrow Pan

32 Original Wide MRA Narrow MRA Exp Wide Pan Narrow Pan

33 OPTIMA Project: Fire Severity Study Hyperspectral data Multispectral data 100% charcoal concentration inside burned area The 100% charcoal area is surrounded by 75% charcoal and 25% vegetation area no 100% charcoal in the inside area no 100% transition area

34 OPTIMA Project: Fire Severity Study ETM image (RGB:5,4,3) ETM image (RGB:5,4,3) 100% charcoal concentration inside burned area The 100% charcoal are is surrounded by 75% charcoal and 25% vegetation area no 100% charcoal in the inside area no 100% transition area

35 OPTIMA Project: Simulator of PRISMA Images Classified orthophoto (SSI 2.5 m, Swath 30 km). Reflectance data cube (spatial res. 7.5 m, spectral res. 2 nm) by using reflectance spectra library. At-sensor radiance by using MODTRAN atmospheric propagation model. Image degraded using PRISMA MTF (30 m) and spectral resolution. Noise effects adding photonic, read out, dark current, and coherent noises Simulated PRISMA Image At-sensor Radiance

36 OPTIMA Project: Procedure for Atmospheric Correction Asphalt Grass Roof Water 1 Asphalt GT Grass GT Roof GT Water GT Reflectance Wavelength (nm) Reflectance (Level L2) from simulated PRISMA Image

37 F L E X Florescence Explorer ESA - AOES Medialab

38 The FLEX (Florescence Explorer) is the 8th ESA Earth Explorer mission. It will make global observations of photosynthesis through the measurement of sun-induced fluorescence (SIF) of the vegetation. Flex will fly in tandem with Sentinel-3.

39 FLEX: Physical principle Fluorescence signal is low (1% to 2%) compared to total reflected signal. For this reason, it is measurable only in Fraunhofer lines of solar irradiance. In this lines, in fact, the Fraunhofer Line Discrimination (FLD) method can obtain sun induced fluorescence (SIF) from the whole reflectance signal. Impinging radiation Reflected Fluorescence Impinging radiation Reflected Fluorescence Target Out of the Fraunhofer line of solar irradiance Target Inside the Fraunhofer line of solar irradiance

40 FLEX: Florescence Imaging Spectrometer (FLORIS) Fluorescence profile

41 FLEX: Florescence Imaging Spectrometer (FLORIS) Hi-resolution narrow-band spectrometer covering two spectral ranges: 677 nm nm spectral range (O 2 -B band) Spectral sampling interval max = 0.1 nm 740 nm nm spectral range (O 2 -A band) Spectral Sampling Interval max = 0.1 nm Wide-band spectrometer 500 nm nm spectral range Spectral sampling Interval max =1 nm Spatial resolution of 150 m binned to 300 m

42 FLEX: Fluorescence Image Simulator for Space Missions (FLISM) The hyperspectral image dataset simulation tool developed for the PRISMA mission is modified to include a fluorescence package in order to produce at-sensor radiance images of fluorescent and not fluorescent targets acquired by a generic spaceborne hyperspectral instrument. The FLISM tool is able to generate subnanometric hyperspectral images of at-sensor radiance including the sun induced fluorescence contribution and it can simulate the measured radiance with or without noise contributions.

43 O 2 -B band band (677 nm 697 nm) O 2 -A band (740 nm nm) At-sensor radiance image Spectral channel 687 nm At-sensor radiance image Spectral channel 760 nm

44 Wide-band spectrometer At-sensor radiance image (R=640 nm; G=540 nm; B=500 nm)

45 LIGHTNING IMAGER

46 LIGHTNING IMAGER

47 Lightning Imager Characteristics FOV Spatial sampling Optical pulse dynamic range (LLp) Optical pulse spectral range 16 diameter shifted northward or 84% of visible Earth disk, including all EUMETSAT member states < 10 Latitude 45 and Sub-satellite Longitude mw/m²/sr ± 0.17 nm Optical channel sensor size Temporal sampling Minimum optical pulse duration Optical pulse size Maximum number of optical pulses in the FOV Instrument Average detection probability (IADP) 1 ms 0.6 ms 10 Km 100 Km circular pulse diameter 25 in 1 ms, 800 in 1 s 70% average over FOV

48 The lightning illuminates a turned off area.

49 The instrument computes pixel by pixel an adaptive background. It uses an adaptive threshold to establish if the recorded value is due to a lightning; the threshold is variable with background: lower in dark areas of the scene, higher in bright ones. It records only the pixels for which the difference between the pixel value and its adaptive background exceeds the threshold.

50 Many of the detected transients (DTs) are not due to lightnings but are caused by noise sources. These DTs are false alarms that have to be individuated and flagged by a 2-steps processing: the first stage is on-board to reduce instrument output data rate and the second stage is on-ground.

51 Pre-processing Noise RTS Particles Ghost Noise Jitter STC Analysis On-ground Filtering Algorithms.

52 A L I S E O Aerospace Leap-frog Imaging Stationary Interferometer for Earth Observation

53 ALISEO Characteristics ALISEO (Aerospace Leap-frog Imaging Stationary Interferometer for Earth Observation) acquires the image as modulated by a pattern of autocorrelation functions of the energy coming from the observed scene. The complete interferogram of any target locations is observed introducing relative source-observer motion (configuration leap-frog), which allows each pixel to be observed while going across the entire pattern of electromagnetic autocorrelation function. Detector Spectral range Spectral resolution Digitalization SNR Detector Silicon CCD 2D-array (1024 x 1024 elements) nm better than nm 12 bit 650 nm

54 ALISEO: Interferogram OPD Due to the absence of entrance slit the device acquires the image of the target superimposed to a fixed pattern of interference fringes.

55 CS Laboratory Prototype

56 Spatial Light Modulator (SLM) Spectrometer Camera

57 The Spatial Light Modulator is the core component of a The Spatial Light Modulator CS system since it physically performs the change of representation domain intrinsic in CS theory.

58 Liquid crystal plate technology High spatial resolution (1920 x 1200 Pixel) Holoeye LC-R 1080 Tiny pixel pitch (8.1 μm) Easy programmable via PC Good radiometric accuracy (8 Bit Grey Levels) Slow frame rate (60 Hz) for real CS applications but adequate for a laboratory prototype

59 Liquid crystal plate technology High spatial resolution (1920 x 1200 Pixel) Tiny pixel pitch (8.1 μm) Easy programmable via PC Good radiometric accuracy (8 Bit Grey Levels) Slow frame rate (60 Hz) for real CS applications but adequate for a laboratory prototype

60 The Liquid Crystal Plate of SLM changes pixel-by-pixel the phase and the polarization state of the reflected light according to the set value. A polarization filter makes apparent the different polarization states of the light reflected by the SLM. The light with a given polarization state can be rejected by turning the polarization filter.

61 First Results

62 The low correlation between the liquid crystal plate pixels and pixels recorded is under investigation.

63 The low correlation between the liquid crystal plate pixels and pixels recorded is under investigation.

64 C A L E T Calorimetric Electron Telescope

65 CALET: Overview The Calorimetric Electron Telescope (CALET) is designed to identify incoming (fully stripped) cosmic nuclei and gammarays and to provide high resolution measurements of their energy. One of the main scientific goals of the CALET mission is to measure the inclusive spectrum of cosmic electrons and positrons in the energy range from few GeV to about 10 TeV. CALET is placed on the International Space Station (ISS) so that it makes long-term observations collecting rare events at high energies, where the fluxes are remarkably low.

66 CALET: Overview Science Objectives Nearby cosmic-ray (CR) sources Dark matter Origin and acceleration of CR Propagation of CR in the Galaxy Solar physics Observation Targets Electron (+ positron) spectrum in multi-tev region Signatures in 10 GeV 10 TeV region of e - + e + and gamma-ray energy spectra p to Fe above several tens of GeV; Ultra Heavy Nuclei B/C ratio up to several TeV /nucleon Electron flux below 10 GeV

67 CALET: Three Detectors CHD Charge Detector IMC Image Calorimeter TASC Total Absorption Calorimeter Function Charge Measurement (Z=1-40) Arrival Direction Particle ID Energy Measurement Particle ID CHD IMC TASC

68 CALET: Expected Results

69 METHODOLOGIES

70 Atmospheric indirect measurements: Data fusion and optimal exploitation of the information content

71 Definition of parameters that evaluate the information content and the quality of atmospheric measurements as a function of altitude O3 Relative Information Distribution MARSCHALS O3 Relative Information Distribution MIPAS-STR Cross-sections of the Relative Information Distribution of MARSCHALS and MIPAS-STR data Blue low information, red higher

72 O3 Relative Information Distribution MARSCHALS

73 O3 Relative Information Distribution MIPAS-STR

74 O3 Relative Information Distribution MSS Data Fusion

75 10,0 (b) Measurement Space Solution 7,5 Development of a new retrieval solution that keeps separated the information of the observations and of the retrieval constraint. Altitude [Km] 5,0 2,5 0, Ozone Volume Mixing Ratio (VMR) Error [%]

76 Altitude [Km] 10,0 (b) Even though MIPAS sounding 7,5 does not observe below 5 Km, the fused VMR error is better 5,0 than MIPAS and IASI. 2,5 0, Ozone Volume Mixing Ratio (VMR) Error [%]

77 Spaceborne microwave radiometers: The SFIM fusion technique to enhance the spatial resolution

78 The Advanced Microwave Remote Scanning Radiometer-Earth Observing System (AMSR-E) has antenna IFOVs of 70x40 km 2 at C-band and of 14x8 Km 2 at Ka-band. The proposed methodology uses Ka-band as high resolution spatial image and injects the details in C-band to obtain a spatial enhanced C-band.

79 The Advanced Microwave Remote Scanning Radiometer-Earth Observing System (AMSR-E) has antenna IFOVs of 70x40 km 2 at C-band and of 14x8 Km 2 at Ka-band. The proposed methodology uses Ka-band as high resolution spatial image and injects the details in C-band to obtain a spatial enhanced C-band.

80 HydroAlgo An algorithm for the retrieval of soil moisture using Remote Scanning Radiometer-Earth Observing System (AMSR-E) data

81 The HydroAlgo utilizes the basic radiative transfer equation which is inverted through an artificial neural networks (ANN) approach. It uses the low-frequency AMSR-E observations at 6.8 GHz (Cband), GHz (X-band) and 18.7 GHz (Ku-band) for estimating soil moisture content and correcting for vegetation effects. This algorithm incorporates as first step the procedure for enhancing the spatial resolution of the lower frequencies previous presented.

82 HydroAlgo Hydrological model

83 HydroAlgo Hydrological model 5 th January th January 2005

84 Space debris population: observation, cataloguing and modelling

85 Observation The orbits of all the un-classified spacecraft are observed by the Space Surveillance Network (SSN). The network is composed by 25 sensors, both radars and optical sensors.

86 Cataloguing The United States Strategic Command catalogues the un-classified spacecraft currently in orbit the Two Line Element (TLE) catalogue. Only the objects larger than 10 cm below a few thousand kms of altitude and about 1 m in higher orbits (up to geostationary) are included. Only less than 10 % of them are operational satellites.

87 Modelling Mathematical models and large numerical Montecarlo codes have been developed to simulate the interplay of all the physical processes involved in the evolution of the debris population. Low Earth Orbit Region: Breakdown of the number of objects larger than 10 cm in the scenario where no future launches are simulated

88 Study and implementation of mitigation measures Active removal of objects from space: theoretical and engineering issues Collision risk evaluation and collision avoidance Traffic management

89 Thank you for your attention Further information: