986
- estimate the radiometric resolution and dynamic
range by calculating statistics and entropy;
- evaluate the spatial resolution and dynamic range
by calculating statistics and entropy;
- evaluate the spatial resolving powers by
o performing an edge analysis,
o estimating point spread functions,
- compare the spectral capabilities,
- analyse the MOMS within scan noise, MOMS and TM
between scan noise and signal-to-noise ratios.
Tables 2, 3, 4 and 5 show different results ob
tained in the testsites.
Fig. 6 shows an enlarged section of the testsite in
Kenya, processed by NASA/GSFC. The MOMS-01 in
frared band is compared to TM band 4. During the
STS-7 mission the MOMS infrared band turned out to
be out of focus. Due to the modular concept the
focal length of the dual lens optics could be
adjusted new for the upcoming STS-11/41-B mission
by mounting a glass plate between optics an arrays.
However, the band still is slightly out of focus.
During the Kenya data take as shown in this example
M0MS-01 was operating in the lower gain mode of the
two, which could be selected by the MOMS mission
specialists at Johnson Space Flight Center in
Houston. Thus the higher spatial resolution of
MOMS versus TM is compensated by a lower dynamic
range (see also Fig. 5).
Complementary to the investigations of the working
group at NASA/GSFC the data comparison was carried
out at the AGF-Working Group Geoscientific Remote
Sensing by employing the following approach:
- display and visual comparison of MOMS 600 and 900
nm bands and TM bands 2+3 (2+3 and average
calculated) and 4,
- MOMS 900/600 nm and TM 4/2+3 ratio processing and
visual comparison,
- MOMS and TM IHS (Intensity, Hue, Saturation)
processing to generate color renditions (see Fig.
7 and 8),
- calculation of histograms (see Fig. 9),
- calculation of scattergrams (see Fig. 10).
Figures 7-10 show the results obtained in the
testsite Atacama desert, Chile.
Figure 7 and 8: MOMS and Thematic Mapper IHS
proessing of a section of the testsite Atacama
desert. MOMS and Thematic Mapper data have been
treated with equal processing methods for comparing
the system's performance for thematic mapping.
Different colors of the same surface features are
due to a different portioning and range of spectral
bands (see also Fig. 5). MOMS features - as given
by the specifications - a higher spatial resolu
tion than TM, which in this specific example is
not being compensated by a lower dynamic range,
because MOMS was switched to gain level 2 (see
also figures 9 and 10). The slight defocussing of
the MOMS infrared band does not afreet the image
quality as the resolution depends on the highest
resolving data used within one processing, which
is the 600 nm visible band.
The sun azimuth angle between both scenes is
nearly perpendicular, thus different geological
features are enhanced.
RESULTS
In the framework of the cooperative effort between
NASA and BMFT MOMS'and Thematic Mapper's system
performances have been compared in terms of spatial,
spectral and radiometric resolving powers and noise
characteristics by NASA.
The evaluations at the AGF-Working Group Geoscienti-
fic Remote Sensing were focussed on preparing
application - oriented computer processings of
both systems under equal conditions and comparing
their applicability for thematic mapping.
The results can be summarized as follows:
- when operating in gain level 1, MOMS has a lower
dynamic range then TM, whereas, when switched to
gain 2, the dynamic ranges are comparable.
- the computer processings prove MOMS to have a
higher spatial resolution, when the exactly
focussed visible band is used directly or
processings including this band. The infrared
band is out of focus, thus point spread function
calculations indicate a lower spatial resolving
power of MOMS versus TM. The spatial resolution
of MOMS decreases from the left to the right
side of one detector array.
- the noise variances in both MOMS and TM are
similar, but since MOMS is a 7-bit system the
percentage of the image radiometric information
containing noise is greater and the SNR of TM
is higher.
- the noise pattern in MOMS images has a vertical
as well as a horizontal structure. The vertical
structure is due to miscalibration of individual
detectors and changing response across detector
arrays. There is also a significant difference in
noise variances between odd and even detectors.
The absolute radiometric correction applied to
the data for the recent evaluations at the AGF-
Working Group improves the pixel grey value
variation within one scan line of maximum 3-4
in raw data to 1 - 2 as could be found by
analyzing image profiles.
- the geometry of the MOMS instrument was found to
be very stable. Although geometrically corrected
TM data has a pixel spacing of 28,5 meters,
subsets of the TM images could be registered to
the MOMS images with an average pixel error of
10 meters.
The results elaborated until now show the two
instruments to be to a certain extend similar in
the spectral bands that were compared. Although
the MOMS scanner has a smaller IF0V, its lower
modulation transfer function versus TM compensates
this advantage especially surface features under
low sun angles or low contrast phenomena have been
observed with the instrument being switched to
gain level 1.
Operating in gain mode 2 an improvement in resolving
power can be stated.
The deficiencies are due to the CCD arrays used
(Reticon CCPD 1728).
For the upcoming M0MS-02 to be tested in space
aboard the D-2 mission actual sensor developments
will be used featuring more pixels per scan line and
a better overall performance (see following chapter).
Optoelectronic arrays have already been tested and
evaluated concerning their applicability for space-
borne remote sensing in the framework of a phase-B
study on Stereo-MOMS, the operational representative
of the MOMS family for the Polar Platform in the
mid-nineties.
TECHNICAL SPECIFICATIONS AND MISSION OBJECTIVES OF
M0MS-02
MOMS-02, as suggested for a space flight aboard the
D-2 mission, will feature the following technical
parameters:
- three nadir looking multispectral bands (560 + 20