218 
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B7. Beijing 2008 
approach here the end of the detector where the quantum 
efficiency is very low; the model is very constrained here and 
adding additional spectral lines would mean only to generate a 
very high error in the center wavelength of this band. 
Table 1: Requirements and Results for APEX Median Radiance 
requirements. 
|nm| 
|nm] 
SSIr 
|nm| 
SSI r 
|nni| 
GSDr 
[m] 
GSDc 
[ml 
380 
387.73 
15 
16.30 
3.65 
3.65 
400 
400.55 
15 
8.80 
3.65 
3.65 
470 
469.85 
10 
8.97 
3.65 
3.65 
500 
500.53 
10 
9.92 
3.65 
3.65 
515 
515.59 
9 
9.59 
3.65 
3.65 
580 
581.77 
5 
4.06 
3.65 
3.65 
650 
650.89 
5 
2.86 
3.65 
3.65 
700 
698.83 
5 
3.51 
3.65 
3.65 
750 
749.61 
5 
4.27 
3.65 
3.65 
780 
781.45 
5 
4.76 
3.65 
3.65 
850 
850.98 
10 
5.89 
3.65 
3.65 
900 
901.81 
10 
6.73 
3.65 
3.65 
940 
937.16 
10 
7.30 
3.65 
3.65 
980 
983.54 
10 
8.04 
3.65 
3.65 
1000 
1000 
10 
8.29 
3.65 
3.65 
Xr 
|nmj 
SNRr 
SNRc 
NeAL R 
NeALc 
F 
380 
314 
716.35 
2.23e-4 
0.98e-4 
- 
400 
681 
692.27 
1.32e-4 
1.30e-4 
- 
470 
484 
924.07 
2.41e-4 
1.27e-4 
- 
500 
737 
897.31 
1.44e-4 
1.18e-4 
- 
515 
901 
879.91 
1.14e-4 
1.17 e-4 
- 
580 
554 
557.14 
1.69e-4 
1.68 e-4 
. 
650 
436 
444.64 
1.87e-4 
1.83 e-4 
- 
700 
313 
450.99 
2.19e-4 
1.52 e-4 
- 
750 
197 
604.96 
4.98e-4 
1.63 e-4 
- 
780 
186 
623.29 
6.44e-4 
1.5 4 e-4 
0.80 
850 
134 
629.56 
1.21 e-3 
1.5 7e-4 
0.61 
900 
138 
590.81 
8.56e-4 
1.99 e-4 
- 
940 
118 
239.08 
3.12e-4 
1.54 e-4 
- 
980 
156 
176.11 
5.38e-4 
4.75 e-4 
- 
1000 
121 
78.89 
6.53e-4 
1.01 e-3 
- 
FHtsr «tosorptövity 
Figure 4: Customized filter for APEX Median Radiance 
Requirements. 
implies that the instrument has the potentiality of performing 
much better than what requested. This is clearly a great 
advantage especially in terms of SNR and noise-equivalent- 
delta-radiance because it will allow (a) to detect a signal much 
higher than its corresponding noise and (b) to distinguish 
between radiance levels that might differ only for a few percent 
of chemical contents. 
Table 2: Binning pattern for the APEX Median Radiance 
requirements. 
1 
2 
3 
4 
5 
6 
7 
8 
1 
35 
132 
160 
172 
214 
243 
258 
34 
50 
140 
167 
178 
215 
243 
258 
9 
10 
11 
12 
13 
14 
15 
271 
278 
291 
299 
304 
310 
312 
271 
278 
291 
299 
304 
310 
312 
Error Budget 
Figure 5: Error Budget For APEX Median Radiance 
Requirements. 
3.2 Dedicated science application: Chlorophyll/Red-Edge 
The optimization algorithm can be applied to specific science 
application and this case study illustrates how this is possible. 
Let’s assume that it is necessary to identify the chlorophyll 
content within leaves with an accuracy of 2%, in the visible- 
near-inffared spectral region. A few reflectance canopy profiles 
of leaves with a content of chlorophyll between 10% and 80% 
are shown in Figure 6. (The PROSPECT 9 model has been used 
in order to generate reflectance curves in step of 2% chlorophyll 
content). When a leave has more than 60% in chlorophyll 
content is very hard to distinguish the variations in reflectance 
because of the high absorption; therefore only curves up to 60% 
of chlorophyll have been considered. 
Nevertheless 80% of the requirements have been met before 
applying any filtering solution. The Figure 5 reports the error 
budget for every requirements; a very positive error means that 
the requirement has been satisfied and, on the other side, 
Figure 6: Canopy Reflectances with variable Leaf Chlorophyll 
Content. The content is indicated in microgramm/cnT.