electro-optical design of imaging payload for a remote

Vol. 2, No. 5, Fall 2009 and Winter 2010pp. 1-14

Electro-Optical Design of Imaging Payload for a Remote Sensing Satellite

E. Peighani-Asl1*, D. Abbasi-Moghadam2, B. Ghafary3 and V. Tabataba-Vakili4

1, 2, 3 and 4. Iran University of Science and Technology *Ave. Farjam, Tehran, IRAN

[email protected]

Remote sensing using small spacecraft arising from multi-objective economic activity problems is getting more and more developed. These satellites require very accurate pointing to specific locations of interest, with high reliability and small latency. The space borne imaging systems always attempted to achieve the highest ground resolution possible with the available technology at the given time. Also mass, volume and power consumption of the spacecrafts and instruments followed the trend to miniaturization. But the most promising prospects for high resolution imaging with remote sensing satellites are connected with passive optical systems, especially push broom systems. In this paper optical system design process is described and different parameters of this process such as MTF, SNR, FOV, aperture diameter, stability and pointing, scanning schemes, detector selection, and target radiance are simulated and analyzed.

Keywords: MTF, SNR, push broom, detector, imaging payload, stability

Introduction1234

High resolution mapping systems follow the trend to smaller Ground Sample Distances (GSD). The increasing number of space borne imaging systems in the last decade (see [1] for more) shows that an increasing number of countries are dealing with space borne technology and that there is an increasing need for mapping systems for different applications [2]. The trend to smaller GSDs was and is supported by the improvements in diverse fields of technology for instance optics, mechanics and materials, electronics, signal processing, communication and navigation. Active micro wave systems, e.g. SAR systems, are an alternative to passive optical mapping systems. They also benefit from the technology improvements. But the most promising prospects for high resolution mapping with small satellites are connected with passive optical systems, especially push-broom systems. High resolution optical systems on small satellites have to overcome a couple of problems. In this paper we consider a GSD of 10 m or less as high resolution, and a satellite with 1000 kg or less mass

1.M.Sc. of Iran University of Science and Technology

(Corresponding Author) 2.PhD. Student in Iran University of Science and Technology 3. Professor in Iran University of Science and Technology 4.Associate Professor in Iran University of Science and

Technology

as a small satellite according to the IAA Study Cost-Effective Earth Observation Missions [3].

Smaller GSD needs larger focal lengths. The physics behind optical systems allows only a restricted number of tricks to overcome the problems of large focal length optics in terms of volume and mass. The size of the focal plane depends on the detector system size and is part of the equation concerning optics volume and mass. The pointing stability is said to be too low using small satellites. What are the requirements and restrictions? The large amounts of date of high resolution imaging systems need to be stored and transmitted using high performance devices. Size, mass and power consumption of those devices increases with increasing data volumes and data rates. High resolution means also to deal with small amounts of energy coming from small ground pixels to be registered in small integration time periods according to the high satellite orbit velocities. So, the question is how far can we go with decreasing the GSD (increasing the ground resolution) using small satellites?

Civil space borne Earth surface mapping started in 1972 with an 80 m GSD provided by ERTS, later renamed to Landsat-1. Nowadays, the GSD approaches 1 m or even less. The electro-optic complex is intended for a panchromatic and spectrozonal survey of Earth surface. Cameras provide reception of optical radiation from Earth surface and

2 / JournalVol. 2,

transformatsignal. Strubased on C

Fig. 1.

Some influence thpoint of viimage quaFunction Mmultiply acomponentwhich mayoptics, CCsystem MT(PSF) of tInverse Fou

Limits o

DiffractioFor most system maand in disc or Airdiffraction important detector pix

2.44.d

whereoptics, anradiation. Ilimited, andetector. Oresolution. average wadesigns sholimited siddetector avof-the-art Coptics wou

l of Space Science No. 5, Fall 2009 a

tion of opticalucture of a ty

CCD, is present

Structure of elec

major featuhe image qualiew. A very eality is to usMTF. Using tall the imagets of a linear syy be based on dCD, electronicsTF. The resuthe system is urier Transform

of electro-op

n of the oper

ay be considerfocus. This di

ry disc. The Ais derived fromparameters w

xel size x:

. 2.44. .f FD

e f is the focal ld is the If x is larger tnd the resolu

Otherwise the Figure (2) s

avelength of ould be near to

de in order to voiding too larCCD detectors uld satisfy this a

e and Technologyand Winter 2010

l image into dypical electro-ted on figure (1

ctro-optical syst

ures are conity from the speffective way se the Moduthe MTF appre quality infystem (or quasdifferent physics …) in orde

ulting Point Spthen compute

m (IFT).

ptical system

rating systemred near diffriffraction causiry disc diamem PSFoptics [4].

which can be

F

length, D is theaverage wavethan d the sysution is deteroptics determ

shows the bor=0.55 m (greo the borderlinget maximum

rge aliasing effwith a pitch o

approach.

digital electrica-optic complex1).

tem complex

nsidered whicpatial resolutioto describe thlation Transferoach, you cafluencing MTsi linear systemcal effects (e. g

er to create thpread Functio

ed applying th

ms designing

ms, the opticafraction limiteses a diffractioeter d caused b. It is one of th

related to th

(1

e aperture of thelength of thstem is detectormined by thines the spatiarderline for aeen). The opticne on the optic

m energy for thffects. For statef 7 m, an f/5.

alx,

chonheeranTFm)g.heonhe

g

aled onbyhehe

1)

heheorhealancscshee-.2

Fig. 2.

AberrAberrathe diadditiodiffracand ch

Table 1

Ima

LatAbeLenAbeSphAbeComAstDisCur

ChromUsualltelescodiffereThis rviolet)also reshows Achromaberratlike apvirtual

SpherCausesfrom acenter This cthan aelimin

EB

Airy disk param(

rations ations are imagiffraction is ional smearinction). It can bhromatic aberra

1. Aberrations thop

age Error

teral Chromatic erration ngth Chromatic erration herical erration matigmatism stortion rvature of Field

matic Aberraly associated wopes. It is the ent wavelengthresults mainly ) around brighteduces lunar a

up more amat doublets tion and morepochromats anlly eliminate it.

rical Aberratis light rays paa mirror) at dito come to foc

causes a star toa sharp point. ate this aberrat

E. Peighani-Asl, DB. Ghafary and V.

meter d as a func( =0.55 m) [8]

ging failures inignored. Aberng effect (ae differentiated

ations (see tabl

hat can be correcptical systems [5

Double/ Schmidt

TeleCass

( )( ) ( )

( )

ation with objectivefailure of a le

hs (colors) toin a faint co

t stars, the planand planetary s speed and

in refractorse expensive, sond those using.

ionassing through ifferent distanccus at differeno be seen as aMost telescop

tion.

D. Abbasi-Moghad Tabataba -Vakili

tion of the f-num

n the image plrrations lead additional tod in monochroe 1) [5].

cted by the three5]

eoptic / segrain

TripleTMA

e lenses of refens to bring ligo a common olored halo (unets and the mo

contrast. It uaperture inc

s help reduceophisticated de

g fluorite lense

a lens (or refces from the ont points on thea blurred disk pes are design

dam,

mber F

lane if to an the

omatic

e Basic

t/

fractor ght of focus.

usually oon. It

usually crease. e this esigns es can

flected optical e axis. rather

ned to

/ 3 Journal of Space Science and TechnologyVol. 2, No. 5, Fall 2009 and Winter 2010


ComaAssociated mainly with parabolic reflector telescopes which affect the off-axis images and are more pronounced near the edges of the field of view. The images seen produce a V-shaped appearance. The faster the focal ratio, the more coma that will be seen near the edge although the center of the field (approximately a circle, which in mm is the square of the focal ratio) will still be coma-free in well-designed and manufactured instruments.

Astigmatism A lens aberration that elongates images which change from a horizontal to a vertical position on opposite sides of the best focus. It is generally associated with poorly made optics or collimation errors.

Distortion The alteration of the original shape (or other characteristic) of an object, image, sound, waveform or other form of information or representation. These distortions are minimized by using symmetric doublets.

Field Curvature Caused by the light rays not all coming to a sharp focus in the same plane. The center of the field may be sharp and in focus but the edges are out of focus and vice versa.

Design of Electro-Optical System Parameters

Optical System Design Process consists of: 1. Determine Instrument Requirements 2. Choose preliminary aperture 3. Determine target radiance 4. Select detector candidates 5. “Optical Link Budget”, SNR considerations 6. Determine Focal Plane architecture and scanning

schemes7. Select F# and telescope/optical train design 8. Complete preliminary design and check MTF 9. Determine ACS requirements

Determine Instrument Requirements The choice of the short-wave (left) border of the panchromatic band is generally conditioned by factual peculiarities of the atmosphere.

Figure (3) shows that the Earth surface illumination when =0.4μm is almost 30 percent, less than when =0.5μm, and the atmospheric fog brightness is 10 – 15 per cent more and atmospheric fog brightness for different Sun angles above horizon. The atmospheric fog influences the quality of images received, as it reduces the general contrast of the image and the signal/noise relation. This leads to the details of low contrast at the image.

Fig. 3. Light exposure of a terrestrial surface ( ), W/m2

Another important reason for the choice of the short-wave border of the panchromatic band 0.5 μm is difficulty in the objective achromatization in the spectral band from 0.4 to 0.5 μm because it requires usage of the special ultraviolet glass which makes optical system much more complicated.

The choice of the long-wave (right) border of the panchromatic band rides on the physical peculiarities of the CCD causing the deterioration of the modulation transmission function in the long-wave band due to generating of the carrier inside of silicon.

The deterioration of the MTF of the multiple-unit linear photo transmitter is caused by the fact that the depth of photon absorption depends on longer waves. Figure (4) shows how the absorption factor depends on wave length and the energy of the photons of pure silicon. The absorption factor is a possibility of photon absorption per unit of length of the way made by light inside the substance. Thus the depth of photon absorption has inverse relationship with the absorption factor. If photons are absorbed outside of the depleted area, the probability of gathering the generated charge by the appropriate sensitive element falls as far as charge moves by diffusion. This leads to the total MTF deterioration.

Fig. 4. Photon absorption in CCD

4 / JournalVol. 2,

DetermineThe aperturhit film. Taperture delarger the brighter (anand imageincreases. A

D 44.2

whereof pixel sidefined Qand 2. For in the opticpixel size. Q=1(see fig

Fig

DetermineThe total atarget at the

totL )(

whereorbit altitudis called atto sensor),(see figurescattered sthe path. Fwith =0.9

Fig. 6. T


e aperture dire is the hole in

The amount oetermines whataperture, the mnd better) the e clarity wilAperture diame

Qdf

e Q is quality ze d to diame = d / DAiry. QQ<1, the resolcs and for Q>As a starting

gure 5) [4].

g. 5. Effect of Va

e target radiaamount of recee orbit altitude

E).(.

e E is the sode (H) from thtmospheric tran, is spectral e 6), and Lpolar and atmoFigure (7) sho [6].

Typical spectralconcrete, soi


ametern a camera thaf light that get a picture wilmore light it cimage will be

ll be appareneter determined

factor. Qualiteter of diffractQ typically varilution is limite1, the resolutiopoint for the

arying Quality F

anceeived spectral e (H) obtains by

pL

lar/lunar specthe earth surfacnsmittance in reflection coefis the sum o

ospheric emitteows target rad

l reflectance curvl, and asphalt [1

at allows light tets through thl look like. Thcollects and the. Greater detant as aperturd by:

(2

ty factor is ratition disk DAiryied between 0.

ed by diffractioon is limited b

design, select

actor [9]

radiance of thy this equation

(3tral radiance ace or target, (the path (targe

fficient of targeof the multipled radiance intdiance intensit

ves for grass, 2]

toheheheail re

2)

ioy ,.5

onbyt

hen:

3)at)

etetletoty

Fig.

DetectThe eabased exampequipmto the the camScanniperformphotodstorage

Ty

Film

PhotoMultipTubeMicroChannPlate

CCD

OpticOne mcamerasignal expresthat earelatio

0i

wspatial

EB

7. Target radian

torsasiest for realon the multie

ple linear CCDment via an axnadir. The lin

mera is situateing (storage) med in seriesdetector line ane and reading b

Table 2. Dete

ype Sensitiv(QE)

~0.01

oplier 0.1 ... 0.

onel 0.1 ... 0.2

0.3 ... 0.

al link budgemore importana for the remoat the zero s

ssed from the ach of the spen signal/noise

2 214

.ob f

dDf hc

where 0iis thel frequency, D


nce intensity in d

lizing is scannelement linear

D, photographixle is directed tnear field (lineaed obliquely to and electrical

s from all thnd then a new begins (see tab

ectors type chara

vity

Non-linear • Dyn. Rang• Detector i• High capa

3• Photocath

(106 ... 10• distortions

2• Photocath

(106 ... 10• no distorti

9

• Dyn. Ranby photon

• Very sensi• Linear sys• High capa

et (Signal to Nnt parameter oote sensing is patial frequengiven lens ap

ectral channels0i [8].

2int

10

. .

S

atm

T EN

e signal to noiis the objectiv


deferent wavelen

ning of the syr photodetectoc plate. In thisto the Earth, uar photodetectthe flight dire

l signals readihe elements o

period of the le 2) [7].

acteristics [7]

Remarks

ge: 1/100 s memory

acity hode, SE – Multi07) sode, SE - Multipli

07) ions nge to 1/1000,Limnoise itive stem (200000 e-/ pacity

Noise Ratio (Sof the electro

the relation oncy 0, that cperture. It is evs may have its

.

dise ratio at theve diameter, f

dam,

ngths

ystemsor, for s case

usually tor) of ection. ing is of the signal

iplier

ier

mited

pixel)

SNR)) o-optic of the

can be vidents own

(4)

e zero is the

ovEit

pevl

cormn

gwacr

af

Electro-Optical D

objective focalvelocity of lighES( ) is the earis the objectivtransmission t

atm( ) is the atsurface reflectphotosensitive efficiency of tvalue (electronlevels of the an

Fig. 8. S

Fig. 9. Sig

The calcucamera providobservation creflectance (clmode with rednominal. Wheshould be prov

Once thegiven. ES( ) iswhen the optiaccount the constraints. real variable psatellite in LEabout 7 km/s. for a ground s

Design of Imaging

l distance, h is ht, d is the elerth surface illumve transmissionthrough the sptmosphere trantance, Tint is

CCD, ( )the photosensitns), correspondnalog-digital co

Signal intensity i

gnal to noise rati

ulation resultsdes high signaconditions. Wlouds, desert), duced storage en the reflecvide within a noe detector is s also given asics is selected

technologica is fixed in m

part is the intEO, the satellit

In other wordsample distanc

g Payload for a Re

Planck’s constement dimensimination by then, f( ) is the pectral channe

nsmission, ( )the storage tis the effectiv

tive CCD and ding to one of onverter(see fig

in deferent sun a

o in deferent sun

s show that pal volume wi

When there it is necessarytime, for exam

tance is normominal storageselected, d ans well as F(f/Dd or designed al and/or th

most cases, so ttegration timee ground track

ds, the dwell tce GSD of 7 m

emote Sensing Sa

tant, c is the ons (d × d), e Sun, ob( )

Light filter el 1 - 2, ) is the Earth time of the ve quantum N0 is signal quantization

gures 8 & 9).

angles

n angles

anchromatic ithin all the is a high y to use the mple 0.5 of mal, survey e time. nd ( ) are

D) and ob( )taking into

he mission that the only tint. For a k velocity is ime is 1 ms m. For high

atellite

resolutionms is too s

)1( mtdwell

Evenfield of vi

1( mIFOVTakin

GSD by geometry about 10-

Thereobstacle: - Use T

with Nand imtechniqmission

- Use thdecreaprojectfor insto thenecess

Noise com

Total

Photoshield + b

2

2

cs

ph

n

n

Electdigitizatio

2darn

Syste2sysn

All oreferencedof the deteelaboratioSignal anany point optical traetc. This because itelectronic

Whermodeled backgroun

2totn

Journal of SVol. 2, No.

n imagers with short for a suff

)10(/) mtdwell

n more severe ew (IFOV).

10(/) IFOVmng both aspeca factor of 10related decrea3.e are two po

TDI (Time DelN stages in ordemprove the SNque is used e. g.ns) he so-called sase the grountion on the su

stance in the ERe satellite veloary dwell time

mponents are:

l noise equivale

on noise: signaackground

2

22

bg

ops

n

nn

tronic noise: deon

22

22det

digro

a

nn

nn

ematic: cross-ta2xtn

of the above nod at the detectoector as the SNn. When evald system noisin the system

ain and filter, apaper referenc

t is an electro-ointerface [8].

re the optics, as thermal

nd radiances ten

2ph nn

Space Science and 5, Fall 2009 and W

GSD of abouficient good sig

10/1 is the influen

100/1)0mcts into accoun0-1 causes a t

ase of energy a

ossibilities to

lay and Integraer to increase t

NR by the fact. in the IKONO

slow-down mod track veloc

urface (those teROS-A1 missiocity in orde

e tdwell.

ent irradiance i

al + optics + fi

2fsn

etector + ampl

2

g

amp

alk + . . . etc:

oise equivalenor. The selectioNR reference pluating the syse system can b

- at the apertuat the detector,ces the surfaceoptical system’

field-stop, andsources, wh

nd to cover the

22sysdar nn

/ TechnologyWinter 2010

ut 1 m, tint < 1gnal and SNR. (

nce of the pix

(nt, reducing ttime related anat the detector

overcome th

ation) technolothe signal N-fotor of N (th

OS and Quick Bi

ode in order city of the liechnique is usion) with resper to obtain t

is expressed in

(

field-stop + col

(

lifier + readout

(

(1

nt irradiances aon of the surfapoint needs somystem SNR, tbe referenced ure, between t, at the amplifie of the detect’s natural optic

d cold shield ahile signal ae entire spectrum

s

5

1/7

(5)xel

6) thendof

his

gy oldhisird

toineed ect the

n

(7)

ld-

(8)

t +

9)

0)

are acemethe

atthe fiertor cal

are andm.

6 / Journal of Space Science and TechnologyVol. 2, No. 5, Fall 2009 and Winter 2010

E. Peighani-Asl, D. Abbasi-Moghadam, B. Ghafary and V. Tabataba -Vakili

Note that a good design requires proper matching of system and cold-shield solid angle. Hence, requires wcs,wsys, with

(11)

Expressions evaluating the noise equivalent irradiance for signal, background, optical train, field-stop, and cold shield follow. Signal:

]/[)()( 22

1

smphdLhcA

fwk

Afk

n

sdd

sd

sdd

ds (12)

Where )(x is the system transmission, f is the noise equivalent bandwidth in Hertz of the band, and is normally given by 1)2(f , effective

integration time (single sample or multiple time delay and integrate samples may be used), dA the detector area,

and dk the noise process scale factor ( 2dk for a photovoltaic and 4dk for photoconductor).

Background:

]/[)()( 22

1

smphdLhcA

fwkn bg

dd

dbg

(13)

Optical train:

]/[)(),,( 22

1

smphdTxTM

hcAfwk

nioi

opiopiopdd

sysdop

(14)

Where each optical element or optical subsystem, i=1,2,3,…,n contributes to the near-field thermal exitance

opM . Note how the remaining optical transmission from the ith component to the detector surface is maintained by dividing the system transmission )(x by the ith component of 0T . In many cases, especially when designing a new system, a priori knowledge of the number, position, temperature, and material properties of the optical elements is unavailable. However, an opn budget can be allocated by approximating the thermal exitance with a single source,

]/[)(

),,( 22

1

smphdTxTM

hcAfk

no

opopopdd

sysdop

(15)

Typically, an KTopop [email protected] will suffice for uncooled optics, where

op is the effective

emissivity, and ][ KTop is the effective temperature.

Field stop:

]/[

)(),,()(

0

2

2

1

smph

dT

xTMhcAwwfk

elsewwif

no

fsfsfsdd

syscsd

syscs

fs

(16) Cold-shield:

]/[

)(),,()(

2

2

1

smph

dTxTM

hcAwfk

no

csscsscsdd

csdcs

(17) The exitance functions, fsop MM , and csM are of

the form

]/[2),,( 22

2

mWdehcdTM Tkhc

b

(18)

where ]/[10381.1 23 KJkb is Boltzmann

constant, is emissivity, and [ ]T K is the temperature. Signal and background radiances,

( )L ds and 2( ) [ / ]bgL d w m sr , from a given

scene that include complex transmission, backscatter, and emission effects can be generated with computer simulation tools such as LOWTRAN or MODTRAN by PCModWin 4.0 software.

When the ADC is matched to the amplifier output then

122NWELL

digNn (19)

where WEELN is the charge well capacity and N number of bits. Ideally, dign is less than the noise floor. This is achieved by selecting a high resolution ADC (large N).

The remaining two noise terms, electronic darn and systematic sysn will not be pursued

any further except within the context of establishing a synthesis-requirement noise budget based on top-level Sensitivity performance requirements ( SNR analysis). Once an darn budget is defined, a technology-dependent feasibility study is performed to identify a candidate technology that meets the darn budget. If no appropriate candidate is found to fulfill the synthesis requirements, the system design is iterated until an acceptable compromise between performance and synthesis is achieved. As a final observation, note that band leakage is a form of systematic noise through spectral cross-talk. Note that correlated noise ( f/1 ,drift etc.) and white noise are not separated in the

n terms. When large enough, this type of correlated noise must be reduced by system calibration

][/4

srDf

wl

sys

/ 7 Journal of Space Science and TechnologyVol. 2, No. 5, Fall 2009 and Winter 2010


in order to avoid imaging artifacts and/or radiometric errors.

In most common remote sensing sensors, namely cameras, the SNR will be in the order of 55 dB, or a ratio of 562: 1. That is, the signal is 562 times greater than the noise signal. At this ratio the noise will be unnoticeable. The following guidelines interpret some ratios of signal-to-noise in terms of the subjective picture quality.

Determine Focal Plane architecture and Scanning SchemesThere are three different scanning systems for acquiring the image (figure 10) [9]:

Fig. 10. Image Acquisition Modes [9]

(1) Whiskbroom imagers are working as electromechanical scanners. On-axis optics or telescopes with scan mirrors sweep from one edge of the swath to the other. The FOV of the scanner can be detected by a single detector or a single-line-detector. Simultaneously the movement of the satellite guarantees the sweeping scan over the earth. This means that the dwell time for each ground cell must be very short at a given IFOV because each scanned line consists of multiple ground cells, which will be detected. A well-known example of whiskbroom imager is AVHRR, Landsat and SeaWiFS.

(2) Push broom scanners: As electronically scanners they use a line of detectors to scan over a two dimensional scene. The number of pixels is equal to the number of ground cells for a given swath. The motion of the satellite provides the scan in along-track-direction, thus, the inverse of the line frequency is equal to the pixel dwell time. By using a two dimensional detector, one dimension can represent the swath width (spatial dimension, y) and the other the spectral range. These imaging spectrometers can be subdivided in Wide Field Imagers (MERIS, ROSIS) and Narrow Field Imagers (HSI, PRISM). By regarding space borne systems these imager-types conduce to high or frequent global coverage.

(3) Staring imagers: these imagers are also electronically scanners. They detect a two dimensional FOV at once. The IFOV along and cross track corresponds to the two dimensions of detector area

array. Two subgroups of staring imagers are Wedge Imaging Spectrometer (WIS) and Time Delay Integration Imagers (TDI). If the incoming light passes a linear wedge filter each row nX of the ground segment is seized by the detector row nX for a determined wavelength. For very high ground resolution and low sensitivity applications, nX rows of the ground can be traced by using a TDI. The light from the scene will be separated by a linear filter for spectral band definition. On the 2D-detector the signal for this line can be read out from multiple rows caused by the forward movement of the sensor. Therefore the sensitivity of a TDI with n rows is n times that of an imager using the push broom principle.

Push broom scanner is selected because these schemes have flowed advantages (see tables 3&4):

- large Swath width and high resolution feasible (depending on pixel number)

- relatively “long” dwell time for each pixel

- high geometric accuracy across the flight direction

- No movable parts

Table 3. Comparison of optical sensor scanning methods [9]

Scanning technique Advantages Disadvantages

Whiskbroom scanner-single detector element

-High uniformity of the response function over the scene

- Relatively simple optics

-short dwell time per pixel

-high bandwidth requirement and time response of the detector

Whiskbroom scanner-multiple detector elements

-Uniformity of the response function over the swath

- Relatively simple optics

-Relatively high bandwidth and time response of the detector

Pushbroom Sensor

-Uniform response function in the along-track direction

- Relatively long dwell time

-High number of pixels per line imager required

-Relatively complex optics

Step-and-Stare Imager withDetector Matrix

-well defined geometry within the image

- long integration time (if motion compensation is performed)

-High number of pixels per matrix imager required

-complex optics required to cover the full image size

-Calibration of fixed pattern noise for each pixel

-Highly complex scanner required if motion compensation is performed.



Table 4. Characteristics of imagers [12]

Characteristic LINE imager Matrix Imager

Pixels 6000-9000 Up to 1024 1024image pixels in

frame transfer mode Photo response nonuniformity

5% 5%

Dark signal nonuniformity

5% 5%

Dynamic range 10000 5000

Select F# and Telescope/Optical Train Design Electro - optical system typesA telescope is an instrument designed for the observation of remote objects. The term usually refers to optical telescopes, but there are telescopes for most of the spectrum of electromagnetic radiation and for other signal types. For example, optical telescope, radio telescopes, X-ray and gamma-ray telescopes [9].

Fig. 11. Basic configurations for Refractive and Reflective optical systems [5]

Electro-optical systems divided into two classifications: reflective systems using specular reflection and refractive systems using optical lens systems (see figure 11). Reflective system is selected because these schemes have flowed advantages:

1. Only one surface of an optical element has to be manufactured with high precision (the mirror)

2. The backside can be lightweight 3. High numerical aperture feasible 4. distortions and optical performance are not

dependent from the wavelength 5. Many different materials are suitable for a mirror

(optical glasses, metals, Zerodur, SiC, etc.)

Infinity F-number or F-stop is defined as: DfF # (20)

where D is the aperture or diameter of the system. F-number indicates the amount that the lens can

open up or close down to let in more or less light,

respectively. The greater the F-number, the less light per unit area reaches the image plane of the system.

Consequently image brightness is proportional to: 222#1 fDFI (21)

In table 5 electro-optical system types advantages and drawbacks are shown.

Table 5. Electro-optical system types advantages and drawbacks [9]

Optics Advantages Drawbacks

Refractive

•Low costs because of spherical lens surfaces •High image quality and strong light signal possible, because of no obscuration by mechanical elements •Inexpensive optic adjustment by circular lenses and axial locations •Diversity of optical lens materials allows spectral corrections in a wide range of applications •Stray light suppression by simple baffle arrangements possible

•No long focal length with high aperture numbers possible •Chromatic aberrations at high spectral band with (> 200nm) have to be corrected

Reflective

•Only one surface of an optical element has to be manufactured with high precision (the mrror) •The backside can be lightweight •High numerical aperture feasible •distortions and optical performance are not dependent from the wavelength •Many different materials are suitable for a mirror (optical glasses, metals, Zerodur, SiC, etc.)

•For small Field of Views only

•More expensive in manufacturing, assembly and integration (higher precision) •Obscuration due to supporting elements for secondary mirror •Higher mass and volume •More stray light

Selection of light filters: Spectral bands of the camera are provided by the

usage of light filters, each one mounted under its own CCD. It was generated taking into account the fact that in the gap between the last lens surface and CCD plane the camera shielding glass (K108 glass) and the CCD shielding glass (K8 glass) were installed. The light filter is sprayed on the mat made of radiation-resistant glass – grade -108. The main optical filter characteristics are listed in table 6 and on the figure (12).

Table 6. Main optical filter characteristics [6]

…nm

med %

med %

0,5 nm

0,1 nm nm nm

med %

max %

520…780 60 5 510…790 500…800 300 1100 0,1 0,5

d

tvqMtloo

t

t

Electro-Optical D

Fig

In table different filter

Table 7. Filt

AdvGLASS FILTERlow impact of inrobustness and l

high transmissibandwidth INTERFERENCany narrow band

spectral fine tun

steep edges characteristics f

ModulatiSome major fthe image quaview. A veryquality is to MTF. Using ththe image qualinear system (on different phorder to create

Considerithe electro-opt

In generathe multiple-un- geometry M

element siz- diffuse MT

silicon absspread;

- MTF deterin the regis

MTF geom

Design of Imaging

. 12. Optical filt

7 the advantaare shown.

ter types advanta

vantages RSnclined entrance ralong-term stability

ion within a lar

CE FILTERS dwidth feasible

ning feasible

in the spectrfeasible

ion Transfefeatures are coality from the y effective wa

use the Moduhe MTF approality influencin(or quasi lineahysical effects the system MTing MTF depentic camera can l there are 3 fanit receiver on MTF determinze and pitch beTF given by spsorption and

rmined by ineffster. metrical multip

g Payload for a Re

ter characteristic

ages and disad

ages and disadva

Disadv

ay limited filtery very limite

characteristicrge fine tuning n

near parallellight requiredsensitive environment

ral additional glfilters necess

er Function onsidered whicspatial resolut

ay to describeulation Transfoach, you can ng MTF compar system) whic (e. g. optics, TF [5]. ndencies the tobe estimated.

actors determinCCD basis:

ned by the phetween them; pectral dependfurther charg

fficiency of cha

plier is as follo

emote Sensing Sa

s

dvantage of

antages [5]

vantages

r assortments d band-pass cs feasible

not possible

l incidence of d

against tal impacts lass blocking sary

(MTF)ch influence ion point of

e the image fer Function multiply all

ponents of a ch may base CCD,…) in

otal MTF for

ning MTF of

hotosensitive

dency of the ge diffusion

arge transfer

ows:

atellite

vMTFg )(

where d ifrequency

If thneighborinto a, the g

vMTFg )(

Figurgeometricwithout apoverlap.

Diffuestimated RETICONCharge Ar

diffusioMTF

wher

)(L

wherspatial chabsorption

Followere accep

Ldr

diff 5010.5

Figurfor monocMTF of inefficienc

(NMTF


dvdv

..)..sin()

is the pitch bety.he apertures ng elements oveometry MTF

dvdv .

..)..sin()

re (13) showsal MTF with ppertures overla

Fig. 13. Geomet

usion constitutheoretically

N technologiesrea (SCA):

on vk e1

1e1

),(

re

2 )2(11

diffLre Ldiff is thharge area dn factor.owing volumespted for furthe

cmcm

4

4

10.0

re (14) shows chromatic radiainefficient chacy volume and

1(exp[),vN


tween elements

of the phverlap and the becomes wors

avav

..)..sin(.

s dependences pitch between ap and a=3

trical MTF of th

uent of the y if we havs diffusion len

diffk

k

k

k

Ldr

Ldr

.1).exp(

)(.1).exp(

2)

he diffusion lepth and k

s of SCA and er calculations

diffusion MTation from 500arge transfer

d number of tran

).2..cos( d

/ TechnologyWinter 2010

(2

s and v is spat

otodiode lineoverlap is equ

se a little bit:

(23

of the receivelements d=7m with apertu

he CCD

MTF can ve typical fngth and Spat

(2

(2

ength, dr is tis the silic

diffusion leng

(2

TF dependenci to 900 nm. depends on tnsfers.

].)). N (27

9

22)

ial

earual

3)

verm

ure

beforial

24)

25)

theon

gth

26)

ies

the

7)

10 / JournalVol. 2,

wherenumber of

Consiregister whhave the in

Fig.

The asurvey is imphotodiodemovement t

MTFmmovement

with sp

movMTF

The fthree storag

Fig. 16. Mdifferent


Fig. 14. Diffusi

e is the transtransfers in thedering catalog

hen two registenefficiency MT

. 15. Transfer in

additional factomage “blur” wits. “Blur” appthat is perpendi

mov-MTF partin focal plane eed Vfp:

fp

fp

VVsin(

following figuge times: 0,

MTFv dependent wavelengths of

maximal stor


ion MTF of the C

sfer inefficience CCD register

gue data (N=40ers are reading

TF given in figu

efficiency MTF

or effecting Mthin the time of pears only aloicular to the CCt conserned during the sign

H

H

TT )

ure (16) gives 0/2 and 0/4.

nce on the spatiaf the incident radrage time d/ Vfp

CCD

cy and N is thr.000 for biphasig, 5 10-5) wure (15).

for CCD

MTF during thsignal storage iong the imag

CD line. with imag

nal storage tim

(28

blur MTF fo

al frequency for diation (with a p).

he

icwe

heinge

geme

8)

or

Tpresen

MTFH

Fig.

whorizoMTF, the ineffic

MTFop

Tpresen

VMTFw

blur w

DetermThere instrumpointinare bemappinnecessgroundimportdistancplatforMTFP

wsinuso

Tsatellit

LMTFw

distanc

EB

Total camera Mnted as follows

)(M

MTFopticH

. 17. Camera MT

where MTFHntal direction,MTFg is the Mdiffusion MT

cient charge tra

co2)(vptics

Total camera Mnted as follows

)( opticV MTFwhere MTFmovwithin the storag

mine ACS Reare many activ

ments, actuatong with high atween arcsecong of the Ear

sary precisionsd control pointtance in orderce and the imarm has three m

.MMTFLMPS

where PS is platidal vibration)

The MTF degrate is

(sin)( xLM cfwhere fx is thece the target


MTF in horizon(figure 17):

).(()(MMTF

MTF

diffusion

gcs

TF at the deferen

is the total , MTFoptics is

MTF of the geoTF, and MTFansfer.

1 )(osvv

vv

MTF in vertica(figure 17):

()( gcs MTFis MTF conserge process.

equirements vities going onrs, and algorit

accuracy. The onds and fractirth’s surface, ds can be corrs. The pointingr to maintain age quality [8]

major componensin.MTFMTFJ

tform stability,

adation due to

).( xLM fa e spatial frequedge moves


ntal direction m

)()

MTF

nt spatial frequen

camera MTthe optical s

metry, MTFdiffF is the MT

2)(1vv

al direction m

)() movMTFrned with the i

n to develop anthms to controobtained accu

ions of degreedeviations fromected using pg stability is ofthe ground sa

]. The MTFPSntsn , J is jitter, and

linear motion

uency, and aLacross the de

dam,

may be

(29)

ncy

TF in ystem fusion is TF of

(30)

may be

(31) image

nd test ol the

uracies s. For m the

precise f more ample of the

(32) d sin is

of the

(33)

LM the etector

pt

aodfm

rit

tfbb

Electro-Optical D

pixel. MTFLMthe motion. Th

t is close approximatelyof aLM on the Mdetector MTFfrequency nomaximum valu

max,x

x

fff

Fig. 18. Influ

For instanspatial frequenrule-of-the-thuimage shift lesthe effect on sy

For MTFthat the jitterfrequency motbe applied. It be described b

e)( JJ fMTFwith J th

shows the influ = 1.x (x – de

Fig. 19. Influecompar

Design of Imaging

only affects thhe distance aLMto the dwel

y MTFg. FigureMTFLM. MTFLFg. The abscormalized to ue,

uence of aLM on M1.5.G

nce, with a detncy of 150 cyumb, when thess than about 2ystem perform

FJ (jitter or ranr is a superpotions so that thesays that manyy a Gaussian d

2exp( 22J f

he rms random uence of on tetector element

ence of s on the rison, MTFD is a

g Payload for a Re

he MTF in the M is v · t. In ll time and e (18) shows thLM with aLM = cissa shows

the system

MTFLM with aLMSD

tector pitch ofyc/mm equals fe linear motion20 % of the d

mance is minimandom motion) osition of diffe central limit y random mov

distribution )2

Jf displacement.

the MTFJ for t size).

MTFJ with s = 0also shown (MT

emote Sensing Sa

direction of many cases MTFLM is

he influence 1 equals the the spatial dependend

(34)

M = 0.1, 1,

f 6.5 m the fi = 1. As a n causes an

detector size, al.is assumed

ferent high-theorem can vements can

(35) . Figure (19) = 0.1.x and

0.1, 1.x. For TFd).

atellite

As a 10% of thminimal eaccuraciesfunctions some caseby simplyduring the

A sinfor systemground res

(sin fMTF2

intt

(J0 –amplitude

Figura of a sinuthumb, whpixel sizeminimal.

FigurMTFg (wit

Fig. 20vibration

Fig. 21.MTFPS,(M( = 0.1 .x)


rule-of-the-thuhe detector sizeeffected. Attituds and stabilizaton micro satell

es, disturbing vy switching ofe relatively shonoidal platformms with largsolution) using

2() sin0x aJf

– Bessel fun of sinus wavere (20) showsusoidal platforhen the amplie, the effect

re (21) showsth detector size

0. Influence of thn on MTFsin wit

MTFD is a

Degradation ofMTFpf) consistin),MTFsin (a = 0


umb, when J ie x, system perfde control systtion to supportlites are under vibrations mayff the active cort imaging pham vibration is tger integrationg

)n xf

nction of orde)the influence rm vibration. Atude is less th

on system

the degradatioe x) due to the

he amplitude a oth a = 0.1, 1 .x. F

also shown (MTF

detector MTFDng of MTFLM (.1 .x). Resulting

/ 1 TechnologyWinter 2010

is less than aboformance is ontems for pointint high resolutidevelopment.

y also be avoidcontrol functioase. to be considern times (low

(3

(3

er zero, asin

of the amplituAs a rule-of-thhan 10 % of t

performance

n of the detectinfluence of

of a sinusoidal For comparison,Fd)

(MTFd) due to(a = 0.2 .x), MTFg MTF is MTFpf

11

outnly ngon In

dedons

redwer

36)

37)

–

ude he-the

is

tor

,

FJf1



MTFPF when for all three components of MTFPFthe rule-of-the-thumb parameters are applied (MTFLMwith aLM = 0.2. x, MTFJ with = 0.1.x, MTFsin with a = 0.1. x). The resulting MTFpf1 equals MTFSR with neglected MTFoptics (MTFoptics = 1).

From an orbit altitude of 600 km, a GSD of 1 m equals an IFOV of 1.7 rad or approximately 1/3 of an arcsec. During the dwell time, the drift shall be less than 20 % of the IFOV resulting in a drift rate of about 2.4 mrad/s or 8 arcmin/s in order to stay in the limit for minimal degradation of the MTF due to drift effects. When using the TDI principle to improve the SNR, for a 96 step TDI the tolerable drift rate becomes even 25

rad/s or about 5 arcsec/s! Required stability for imaging satellite is:

int

1TH

RStability (38)

where R and H are resolution and orbit altitude respectively. Amount of for the different resolution is shown in table (8).

The accuracy of pointing is:

HS1.0tan 1 (39)

where S is the swath width.

Table 8. amount in the different resolution

Resolution (m)

>500 0.5 Between 500 and 100 0.33 Between 100 and 30 0.15 Between 30 and 5 0.1

<5 0.01

Data Volume & Transmission Rate The data rate required for observation payloads depends on the resolution, coverage area, amplitude accuracy, and number of sensors. When the satellite is in low-Earth orbit, the satellite motion itself allows easy scanning of the Earth in the orbit plane. A separate mechanism in the sensor scans perpendicular to the orbit plane. The sensor generates an image composed of minimum resolvable elements called pixels. If the resolution element's diameter on the ground is d meters, directly below the satellite, the pixel size is hd / radians, where h is the satellite's altitude. The width of the sensor's scan angle, perpendicular to the satellite orbit, is x also in radians [9].

The data rate, DR, generated by the sensor is:

sbitsqdhsbVDR Nx /2

(40)

Where NV : satellite's ground-track velocity, x : width

of the sensor's scan, perpendicular to satellite's orbit (radians), d : minimum diameter of pixel image

projected on the ground, h : satellite's altitude, s :number of samples per pixel (typically 1 to 2), b :number of bits per sample ( b2 amplitude levels), q :frame efficiency, fraction of time for data transmission (typically 0.90 to 0.95).

Mass, Volume, Power Consumption

Microelectronics Since the launch of Landsat-1 in 1972, the progress in microelectronics has enabled more sophisticated instrument designs. The developments for the MESUR Network Mission may serve as an example, how much microelectronics technology may influence the overall mission design. The MESUR (Mars Environmental Survey) Network Mission concept consisted of up to 16 small spacecraft (that time planned to be launched in 2001). As often in extraterrestrial missions, there was a pressure to miniaturization by need. Reference mission was the MESUR Pathfinder Mission, one of the first missions under NASA’s Discovery program of smaller, low-cost missions to be launched 1997.

In [3] the benefits have been assessed which may occur when the electronics technology used in the MESUR Pathfinder mission is replaced by advanced microelectronics technology. The MESUR Network study team found out that advanced microelectronics packaging technologies could be applied to the implementation of subsystem functions for

- The Attitude and Information Management System AIMS

- The Radio Frequency Subsystem RF - The Power and Pyro Subsystem PP.

As a result, a factor of three or better reduction in mass, volume, and power consumption were projected relative to the MESUR Pathfinder baseline (see table 9).

Table 9. Projected total reduction in mass, volume, and power consumption for MESUR Network in comparison to MESUR

Pathfinder [10]

Pathfinder Network Next Reduction Fractional

Mass 47 Kg 11 Kg 36 Kg 4.3x

Volume 46 dm3 6.5 dm3 39.5 dm3 4.1x

Power 74 W 26 W 49 W 2.9x

The key to realize these reductions lies in the utilization of industry- based advanced microelectronics packaging technologies, including:

- Multichip module (MCM) technology - Three-dimensional MCM stacking - Die stacking for memory.

The leverage of these reductions to the spacecraft is obvious. The advanced microelectronics packaging technologies have been widely used for instance in a joint NASA/DLR study for the ROSETTA lander carrying among other cameras a stereo camera with 10 mm GSD

cbi

a

P

tdfpOi

be(

mff

i

Ticmo

Electro-Optical D

[5] and in a joconcept for plabeen remarkablin a very smallswath width ofand with a wei12.5 Watts incl

Pixel size influsize of the detthe ground pixdetector elemefigure 22). Apixel size of 7Of course witintegrated. If tsufficient to obe applied orenlarging the (should not be

Fig. 22. Depennecessary focal

1 m fr

Impact ofmass of opticsfocal length afield size deterstaggered line

- Detector l- Image fie- Focal leng- the optics

many lineline pairs Staggered

in the SPOT-5

The focal lengis determined contradiction tmass and voluof optic system

Design of Imaging

oint DLR/NASanetary explorale. The latter col stereo cameraf 250 km from ight of 2 kg anluding a 1Gbit m

Detecuence - For mtector is projecxel size to be ents x the shors an example

m results in ath smaller detethe sensitivity oobtain the necer a so called

dwell time used in stereo

ndence between length f for a gi

from an orbit alti

f staggered cons depends sign

and the aperturrmined by the arrays, the follline length is h

eld area is reducgth is halved s need to be oe pairs per milper millimeter

d CCD-line arr mission came

Optigth of high resoby the physicsto a small sat

ume. The progrms enables now

g Payload for a Re

SA three-line station [6]. The oncept resulteda for a GSD ofan orbit altitude

nd a power conmass memory [

ctormapping purposcted via the foc

obtained, the rter the focal le, the stat-of-ta focal length oector sizes lesof the pixel ele

essary SNR, Tslow-down mto the sufficimaging) [11]

detector elemeniven ground pixeitude of 600 km

nfigurations - Vnificantly not re, but also ondetector extenslowing effects halved ced to one qua

of high quality llimeter with renecessary for thrays are used ras HRS.

ics olution space bs laws and seeellite concept ress in productw the utilizatio

emote Sensing Sa

tereo camera effects have

d for instance f 20 m and a e of 250 km, nsumption of 10].

ses the pixel cal length to smaller the

length f (see he-art CCD of f = 4.2 m. ss energy is ement is not

TDI needs to mode allows cient extent .

nt size x and el size of X = [8]

Volumes and only on the n the image sions. Using occur:

arter

for twice as espect to the he pixel size. for instance

born sensors ems to be in in terms of

tion and test on of highly

atellite

efficient lfor space a

- Use of a- Use of f

(e. g. TM- Use of s

Even compatibleof the lenimagers isenvelope f

TheperformedconstructiDST, [2].Newton tplaced viaalready inrequiremelaunch. TDobson wastronomesensing pupushes thmaximumsize of abo

DesigThe propomain char

1. Resolthe alkm fo

2. Cover 200

3. Earth -0.9 The d

listed in TTable 10

Orbit altiFocal disLens apeSpectral Spatial reSwath, kNumber channel (EquivaleIntegratioTotal Mline/mm)Signal toVideo daStorage cSpeed ofData tranEntrance


ow-mass and applications. E

aspheric lenses folded arrangemMA) ophisticated caif you can d

e with a microns system fors a problem if for piggy-back l

Technical Ud a study conon approach: . the core eletelescope. Tha four 2.1 m bn orbit. In orents it is foldedThis type of twas originally iers. To increaurposes, a “Bahe focal ratio

m possible magout 1m from a

gning Resulosed system s

racteristics: lution (GSD) ltitude of 700 kor operation ofrage area durin00 km

surface remotem for panchro

designing resuTable (10). 0. The designing

itude, km stance of objectiveerture, D/F bands, μm esolution, (ground

kmof equivalent CC

(14x14 m) ent CCD element son Time, ms

MTF at the Nyq), minimum o noise ratio, minimata quanting capaccapacity, Gbit f data saving, Mbitnsfer rate to radio ce pupil diameter, m


low-volume opExamples are

in refractive tements for refle

atadioptric teledesign a camer-satellite spacer high resolutiyou think of t

launch opportunUniversity of ncerning an inthe Dobson Sement of DSTe secondary booms when trder to fulfilld to minimal stelescopes callinvented by amase the resoluarlow lens” wit

up to f/12.5gnification and700 km orbit.

lt of Imaginshould provid

of Earth surfakm: 50 m with

f panchromatic ng Earth surfac

e sensing in spomatic camera;ult of a panchro

result of a panch

es, mm

d pixel projection)

CD elements in e

size, μm

quist frequency(3

mum ity, bit

t/schannel, Kbit/s

mm

/ 1 TechnologyWinter 2010

ptical telescop

elescopes ective telescop

escopes. ra having weigcraft, the volumion space borthe restricted sinities. Berlin recent

nteresting optiSpace TelescopT is a 20´´ fmirror will

the spacecraft l micro satellspace during tled truss desimbitious amateution for remoth a factor of 25 which assurd a ground pix

ng Payloadde the followin

ace survey froh swath width

camera; ce survey: 50 k

pectral bands: 0;omatic camera

hromatic camera

700199.6 1:5.4

0.5 – 0.9 m 50

50 ach 1024

14 147.2

35.6 0.35

100 8

2.0 1.115 1330 37.0

13

pes

pes

ght merneize

tlyics pe,f/5 be is

ite thegn

eur ote2.5res xel

ng

om50

km

0.5

is

a



Table 11. The designing result of a panchromatic camera, Continuance

Main wavelength, m 0.5893 Angular field of view (2 ), degrees 12 Distortion, % < 0.1 Integral light transmission factor >80% Field of view illumination change, % < 5 Pointing, degree < 0.4 Stability, degree/second < 0.08

Conclusion

This paper showed the problems connected with the imaging payload design. In this context the paper deals with such important parameters for imaging payload for remote sensing satellites like spatial resolution, MTF, signal, SNR, pointing accuracy and stability.

The systematical design and performance evaluation of the Imaging Payload for the remote sensing satellite technology has been described. It was done based on top-level customer system performance requirements and proposed approach for this purpose is based on SNR (detection) and MTF (resolution) analysis. The Imaging Payload was designed as a push broom scanner flying in a sun-synchronous polar orbit of 700 Km altitude and resolution of 50 meter.

References [1] Jacobsen, K., High Resolution Satellite Imaging

Systems – an Overview, PFG 6/2005, 487-496. [2] Segert, T., Danziger, B. and Geithner, M., “The

Dobson Space Telescope – A time shared Telescope for NEO and Earth Observation,” Paper IAA-B4-0605 of the 4th International IAA Symposium on Small Satellites for Earth Observation, April 7-11, 2003, Berlin, German.

[3] Kramer, H. J., Observation of the Earth and its Environment – Survey of Missions and Sensors, 4th

edition, Springer, Heidelberg, New York, 2002.

[4] Jahn, H. and Reulke, R., Systemtheoretische Grundlagen optoelektronischer Sensoren, Akademie Verlag, Berlin, 1995.

[5] Smith, W. J., Modern Optical Engineering, 2nd

Edition, McGraw-Hill, 1996.

[6] International Study on Cost-Effective Earth Observation Missions, Edited Rainer Sandau, German Aerospace Center (DLR), Taylor & Francis the Netherlands, in print, 2006.

[7] Konecny, G., Geoinformation: Remote Sensing, Photogrammetry and Geographic Information Systems, Taylor & Francis, London and New York, 2003, 248 p.

[8] Holst, G. C., CCD Arrays, Cameras, and Displays, SPIE Optical Engineering Press,Bellingam, 1998.

[9] Wertz J. R. and Larson, W. J., Space Mission Analysis and Design, 3rd Edition, Microcosm Press, ElSegundo, California, 1999, 969 p.

[10] Alkalai, L. And Davidson, J. (Eds.), MESUR Network – Integrated Microelectronics Study. Final Report, JPL D-11192, January 7, 1994.

[11] Sandau, R., Hilbert, S., Venus, H., Fang, W.-C. and Alkalai, L., “A Three-Line Stereo Camera Concept for Planetary Exploration”, Paper IAA-L-0902 of the 2th International Conference on Low-Cost Planetary Missions, April 16-19, 1996, Laurel, Maryland, USA.

[12] Holst, G. C., Electro-Optical Imaging System Performance, SPIE Optical Engineering Press, 1995.

electro-optical design of imaging payload for a remote

Documents