338 
Despite CHRIS is a hyperspectral sensor, of which the 
spectrum resolution is above lOnm, the band design of it is 
rougher, compared with the spectrum width of Handhold Field 
ASD spectrometer (spectrum resolution is above lnm).So there 
is a coupling between the bands chosen of CHRIS and the 
retrieval algorithm based on the data Handhold Field ASD 
measured, i.e. the band design of CHRIS is not satisfied the 
retrieval algorithm accurately. As a matter of fact, we choose 
band 6(the mid band is 561nm) corresponding to 566nm, band 
9(the mid band is 622nm) corresponding to 617nm, band 
10(the mid band is 656nm) corresponding to 652nm, and apply 
the retrieval algorithm on the CHRIS images which we obtain 
on April, 1 st , 2006.The distribution map of chlorophyll 
concentration retrieved illustrated as follows (See Fig 7). 
Figure 7.The distribution of chlorophyll concentration in the 
Three Gorges Dam 
As we can see, the distribution of chlorophyll concentration in 
the Three Gorges Dam area is obvious. The concentration of 
chlorophyll in the centre of the dam is relatively lower, 
compared to the riverine is higher. Specifically, in the upper 
course of the dam ,the quality in the midstream is better than in 
the riverine; and the range of chlorophyll concentration is 
between 2.3~2.6 mg/m 3 and 2.7~2.9 mg/m 3 respectively; In 
the lower reaches of the dam, the quality of water is depressed 
near the dam; in the “Y” area, the velocity of water flow is 
slow and the concentration is up to 2.9 mg/m 3 . Generally, the 
concentration of chlorophyll in the Three Gorges Dam is low 
and the water in it is limpid. 
Table 8 shows the results of the comparison between the 
retrieved values from CHRIS images and the measured values 
of chlorophyll. The minimum retrieval error is 16.16%, the 
maximum retrieval error is 33.23%, and the average error is 
24.24%. It is promising that CHRIS can be applied in water 
quality remote sensing monitor of inland water. 
Samples 
Calculation 
(mg/m3) 
Measurement 
(mg/m3) 
Error 
% 
Sample 1 
3.52 
2.48 
29.44 
Sample2 
2.89 
2.42 
16.16 
Sample3 
3.01 
2.43 
19.17 
Sample4 
3.24 
2.44 
24.55 
Sample5 
3.15 
2.44 
22.63 
Sample6 
3.42 
2.39 
30.07 
Sample7 
2.76 
2.30 
16.63 
Sample8 
3.10 
2.41 
19.32 
Sample9 
3.78 
2.52 
33.24 
Sample 18 
3.58 
2.46 
31.21 
Table 8 .The comparison between the retrieved and the 
measured values of chlorophyll 
4. CONCLUSIONS 
The CHRIS/PROBA system provides high spatial resolution 
hyperspectral/mutiangular data, what constitutes a new 
generation of remote sensing information to be processed and 
exploited. The results show that CHRIS/PROBA images can be 
applied in water quality remote sensing monitoring of the Three 
Gorge Dam area. Specifically, the combination of band 6,9,10 
can be adopted in remote sensing monitor of chlorophyll and 
the retrieval precision is better. 
It is significative and estimable to apply CHRIS/PROBA 
images in water quality remote sensing monitor of inland water. 
As the number of samples which can be used is limited, the 
validation of the model has not done yet, so the model is 
restrictive on a certain extent. In addition, the statistic model 
itself also influences the precision and the applicability. 
Therefore, with the spectrum absorption coefficient, scatter 
coefficient to establish the Bio-Optical model is the key to 
amend models and improve the precision of remote sensing 
monitor. 
REFERENCE 
AHN Yu-Hwan, MOON Jeong-Eun, SONIA Gallegos. 
Development of suspended particulate matter algorithms for 
ocean colour remote sensing. Korean Journal of Remote Sense, 
2001, 17(4):285-295 
Alessandro Barducci, Donatella Guzzi, Paolo Marcoionnni, 
Ivan Pippi. CHRIS-PROBA Performance Evaluation: 
Signal-to-Noise Ratio, Instrument Efficiency and Data Quality 
form Acquisitions over San Rescore Test sitt.Proc Of the 
3 rd .ESA CHRIS/PROBA workshop 31-23 March,2005,Frascati. 
Dong Guang-xiang, Zhang Ji-xian, Liu Zheng-jun.A 
Comparison of several destriping methods for CHRIS/PROBA 
data. Remote Sensing Information, 2006, 6:36-39 
http ://www. earth. esa. int/proba/HDFclean v2. html 
http://www.people.com.cn/GB/jinji/222/10814/10824/10836/20 
030528/1002716.html 
Li Su-ju, Wang Xue-jun. The spectral features analysis and 
quantitative Remote sensing Advances of Inland water Quality 
Parameters. Geography and Territorial Research, 2002, 18(2): 
26-30 
Ma 
algo 
data 
PA 
cone 
trans 
Rem 
Shu 
4(1): 
Ted 
a ki 
valid 
work 
Thiei 
conte 
IRS- 
Remt