The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008 
averaged for each spatial and spectral pixel, corrected for dark 
values, and then compared to the reference values for the 
corresponding light source and illumination levels. The detailed 
analysis allows checking for sensitivity (slope) of all pixels, 
changes in pixel-related non-uniformity, and bad pixel 
recognition. Since the analysis always assesses the complete 
channel from the light source to the detector array, in general 
only a summarized characterization is possible and not a 
separated one for each single system component. In some cases 
such a single system component can be analyzed by 
investigating different measurements. If deviations from 
reference values are found, a more detailed analysis is 
necessary, possibly in connection with additional measurements, 
i.e. Sun calibrations. In the case that changes of a light source 
are found, which are stable and can be cross-proven from other 
measurements, the reference data for the corresponding internal 
lamp in the calibration product have to be updated. 
2.3 Sun Calibration Measurements 
The measurement of the Sun through a full aperture diffuser 
provides the only feasible opportunity for absolute spectral 
calibration of the HSI after launch. All other measurements 
only allow for a relative assessment. Assuming normal 
functioning and behavior of the instrument, a rather low 
frequency of Sun calibrations measurements is planned. Each 
Sun calibration measurement is planned to be accompanied by 
internal lamps measurements. To achieve the necessary 
accuracy, especially the geometry between the full aperture 
diffuser and the Sun during the measurement itself has to be 
known. Based on this information the part of read-outs with 
correct incident angle of the Sun can be selected. These 
measurements for all spatial and spectral pixels are used for 
averaging and calibration analysis. As long as all instrument’s 
parameters are stable, the Sun measurements should be 
reproduced with a maximum deviation of 1%. If larger 
deviations are found, a more detailed analysis is necessary, 
including cross-checks with internal lamp measurements to 
verify causes and consequences of observed changes. If the 
changes can be proven, a corresponding update of the 
calibration product will be the consequence. 
3. PROCESSING CHAIN 
The design of the EnMAP processing chain is based on the 
experience with a fully automated and ISO 9001-2000 certified 
processing chain for airborne hyperspectral data (Bachmann, M. 
et al., 2007; Habermeyer, M. et al., 2005). Similar to this 
processor chain, the newly developed EnMAP processors will 
include system calibration, parametric geo-coding, atmospheric 
correction, and assessment of data quality. 
EnMAP level 0 products (raw data) will be long-term archived, 
while EnMAP level 1 (systematically and radiometrically 
corrected data), 2a (geometrically corrected data), 2b 
(atmospherically corrected data without geometric correction), 
and 2 (atmospherically corrected data with geometric correction) 
products will be processed and delivered to the user without 
archiving. The EnMAP level 0 processor mainly collects data 
from the different sources. Beside the datatake itself, it derives 
additional information, e.g. the quality of the acquired data. The 
EnMAP level 1 processor corrects the hyperspectral image for 
known effects, e.g. radiometric non-uniformities, and converts 
the system corrected data to physical at-sensor radiance values 
based on the corresponding valid calibration values and dark 
measurements (see section 2). The EnMAP level 2a processor 
creates orthoimages by direct georeferencing utilizing 
navigation data and an adequate digital elevation model. The 
extraction of ground control points from existing reference 
images by image matching techniques - if suitable reference 
images are available - serve to improve the line-of-sight vector 
and therefore to increase the geometric accuracy of the 
orthoimages. The EnMAP level 2b processor converts the 
physical at-sensor radiance values to ground reflectance values 
separately for land and water applications. This includes the 
estimation of the aerosol optical thickness and the columnar 
water vapor. Figure 2. illustrates this part (“Processing Chain, 
Level > 0” of Figure 1.) of the processing chain. 
( ) Level 0 Product (parts) 
Processing Chain 
Level 1 
J Orbit and Attitude Products 
(Systematic and Radiometric 
Correction) 
I 
Level 1 Product 
Level 1 Product 
t 
Processing ¿hain 
Level 2a 
(Geometric Correction) 
Level 2a Product 
Processing Chain || 
Leve! 2b 
(Atmospheric Corrector) || 
Level 2b Product 
I 
Level 2 Product 
I 
Figure 2. Processing Chain 
Two other spacebome hyperspectral instruments are currently 
operated for civil Earth observation. These are the technical 
demonstration missions Hyperion on EO-1 by NASA/USGS 
(launched on November 21, 2000) and Chris on Proba by ESA 
(launched on October 22, 2001). While Hyperion/EO-1 
distributes level 1 and level 2a products, Chris/Proba provides 
level 1 products only. 
3.1 Transcription 
The EnMAP level 0 processor mainly collects information from 
the different data streams, extracts and interprets information, 
and evaluates and derives additional information, creating the 
EnMAP level 0 product. This EnMAP level 0 product 
comprises: Image tiles (for Earth, Sun, and deep space 
measurements), bad pixel/line/channel information, quicklooks, 
cloud and haze information, water-land information, metadata, 
and dark measurements (see section 2). 
3.2 Systematic and Radiometric Correction 
The EnMAP level 1 processor corrects the HSI image data for 
known systematic effects like odd-even and non-uniformity, 
and then the processor converts this system corrected HSI 
image data to physical at-sensor radiance values based on the 
corresponding valid calibration and dark values. The EnMAP 
level 1 product comprises: Image, bad pixel/line/channel mask, 
cloud and haze mask, water-land mask, metadata, processed 
orbit and attitude, and dark value information. Figure 3. 
illustrates this part (“Processing Chain, Level 1” of Figure 2.) of 
the processing chain. 
Processing Chain 
Level 1 
(Systematic and Radiometric 
Correction) 
—I— 
Figure 3. Systematic and Radiometric Correction