To accurately assess the area of land cover in hill land, we integrated DEM data and remote sensing image in Lihe River
Valley, China. Firstly, the DEM data was combined into decision tree to increase the accuracy of land cover
classification. Secondly, a slope corrected model was built to transfer the projected area to surface area by DEM data. At
last, the area of different land cover was calculated and the dynamic of land cover in Lihe River Valley were analyzed
from 1998 to 2003. The results show that: the area of forestland increased more than 10% by the slope corrected model,
that indicates the area correcting is very important for hill land; the accuracy of classification especially for forestland
and garden plot is enhanced by integrating of DEM data. It can be greater than 85%. The indexes of land use extent were
266.2 in 1998, 273.1 in 2001, and 276.7 in 2003. The change rates of land use extent were 2.59 during 1998 to 2001 and
1.34 during 2001 to 2003.
Systematic mapping and monitoring of wetland landscape are of fundamental importance for wetland development and
management. To accurately classify wetland in Yancheng coastal wetland, ground investigation was conducted in 2006.
Integrated with ground investigation, the wetland was classified into 8 categories such as Spartina alterniflora Loisel,
Farm land, Phragmites Australis, Artemisia halodendron Turcz, Bare beach, Salt field, Fish & shrimp pond, and Sea
water. A total of three decision trees were successfully produced. The first represented broad divisions of vegetation (in
fact, at this stage, it just can be called vegetated cover like) and non-vegetation, and the second two represented more
detailed vegetation classes and non-vegetation classes. To construct the decision trees, NDVI and principal component
analysis were used as the evaluation factors. The thresholds were built combining with ground investigation and spectral
property. Firstly, almost all kinds of vegetable were divided out of non-vegetation by NDVI. Secondly, the different
species of vegetation were distinguished and some vegetated cover like was eliminated out of vegetation. Phragmites
Australis belt, Artemisia halodendron Turcz belt, Spartina alterniflora Loisel belt and bare beach belt were distributed
regularly from land to sea.
In this paper, we want to search for the hyperspectral remote sensing bands most sensitive to chlorophyll-a concentration
in different trophic states. Through repeated measurements in Taihu Lake, a large quantity of hyperspectral reflectance
data and chlorophyll-a concentration data of lake were obtained from June to September of 2004 and 2005. The eligible
hyperspectral data were analyzed to calculate remote sensing reflectance of water in Taihu Lake, and the data of
chlorophyll-a concentration obtained from laboratory analysis were used to calculate Trophic State Index. According to
the actual condition of Taihu Lake, the hyperspectral data were divided into three groups: mesotropher, eutropher and
hyper eutropher. In each trophic state, chlorophyll-a concentration was then regressed against to identify the most
sensitive hyperspectral bands. From the established regression models, we can get the conclusion: chlorophyll-a
concentration is correspondingly lower under mesotrophic state, badly influenced by suspended matter, the spectral
feature of chlorophyll-a is not evident, and the accuracy of regression model in this trophic state is not so satisfactory; in
eutrophic state, differential algorithm has better retrieval result, the linear model based on this algorithm has the best
estimation result; under hyper eutrophic state, the estimation accuracy is higher than the other two states as a whole. The
fitting precision is the highest using the band ratio R719/R670 as independent variable in the quadratic equation model.
Based on TM (ETM) data and in-situ measurements of chlorophyll-a concentration (Chl-a) in Lake
Taihu, analysis was conducted to decide the correlation between Chl-a and the ratios of different
reflectance corrected by the 6S model. The results show Chl-a is closely related to TM3/(TM1+TM4)
and the inverse model to infer Chl-a in Lake Taihu can be written as
Ln(Chl-a)=-9.247*(TM1+TM4)/(TM2+TM3) -27.903*TM3/(TM1+TM4) +24.518. However, the accuracy of this model can not be assured due to the complexity of spectral reflectance strongly dependant on water quality in Lake Taihu. Thus we developed a further 2-layer BP neural net model based on 4 input nodes, 7 hide nodes and 1 output node to for calculating Chl-a in the lake. The derived results reveal that the BP model has much higher accuracy than the linear model. A test was made based on 16 samples and suggests that the maximum relative error (RE) of BP model was only 35.43%. Of all the samples, 15 ones had a RE of less than 30% from the BP model.. However, there were only 3 samples with RE less than 30% from the results derived from the linear model. The comparison shows that the BP model has high availability for inferring Chl-a of surface water having complex spectral reflectance.
Remote sensing technique is soundly used in water quality monitoring since it can receive area radiation information
at the same time. But more than 80% radiance detected by sensors at the top of the atmosphere is contributed by
atmosphere not directly by water body. Water radiance information is seriously confused by atmospheric molecular and
aerosol scattering and absorption. A slight bias of evaluation for atmospheric influence can deduce large error for water
quality evaluation. To inverse water composition accurately we have to separate water and air information firstly. In this
paper, we studied on atmospheric correction methods for inland water such as Taihu Lake. Landsat-5 TM image was
corrected based on Gordon atmospheric correction model. And two kinds of data were used to calculate Raleigh
scattering, aerosol scattering and radiative transmission above Taihu Lake. Meanwhile, the influence of ozone and white
cap were revised. One kind of data was synchronization meteorology data, and the other one was synchronization
MODIS image. At last, remote sensing reflectance was retrieved from the TM image. The effect of different methods
was analyzed using in situ measured water surface spectra. The result indicates that measured and estimated remote
sensing reflectance were close for both methods. Compared to the method of using MODIS image, the method of using
synchronization meteorology is more accurate. And the bias is close to inland water error criterion accepted by water
quality inversing. It shows that this method is suitable for Taihu Lake atmospheric correction for TM image.
Inherent optical property is an important part of water optical properties, and is the foundation of water color
analytical model establishment. Through quantity filter technology (QFT) and backscattering meter BB9 (WETlabs Inc),
absorption coefficients of <i>CDOM</i>, total suspended minerals and backscattering coefficients of total suspended minerals
had been observed in Meiliang Bay of Taihu lake at summer and winter respectively. After analyzing the spectral
characteristics of absorption and backscattering coefficients, the differences between two seasons had been illustrated
adequately, and the reasons for the phenomena, which are related to the changes of water quality coefficient, had also
been explained. So water environment states can be reflected by inherent optical properties. In addition, the relationship
models between backscattering coefficients and suspended particle concentrations had been established, which can
support coefficients for analytical models.
Light is very important for water ecosystem, it was attenuated because of absorption and scatter caused by suspend
sediment, chlorophyll and colored dissoluble organic matter(CDOM) in water. Water remote sensing reflectance is an
important parameter for ocean color remote sensing, which has good relation with water quality. Measure underwater
spectrum has significance application value for water ecosystem research and water quality retrieval. This paper
introduced the TriOS spectrum measurement system firstly, then use the in situ data collected by the system in Taihu
Lake to retrieval suspended sediment(SS) concentration. First, underwater remote sensing reflectance(Rrs(0<sup>-</sup>)) was
calculated by using the underwater spectra, then transform the Rrs(0<sup>-</sup>) into Rrs(0<sup>+</sup>) based on water-air interface
transmission model, after that analysis the relativity between SS concentration and Rrs(0<sup>+</sup>), and finally suspend sediment
retrieval models were developed. Compare different models, in order to select an optimal model for retrieving SS
concentration. The result indicates that: Rrs(0<sup>+</sup>) has good relationship with the logarithm of SS concentration(Ln(SS)). In
spectral region 500nm-600nm, there present negative correlation, and present well positive correlation in spectral region
620nm-882nm. The maximal correlation coefficient locates at band 743nm. Take Rrs(522), Rrs(743), Rrs(815) as
independent variable to build SS concentration retrieval models, the retrieval results showed that logarithm model is
better than other models, it can satisfied the require of practical application to certain extent.
Chlorophyll is a very important parameter for lake water quality evaluation. Its concentration varies seriously with
different season. The chlorophyll-a concentration inversing models in different season were studied using different
temporal TM images. The models were built in 3 steps: Firstly, 10 images were selected according to the principle of
almost synchronously with in situ measurement; secondly, remote sensing images were preprocessed. Atmospheric
corrections were carried out use 6S model, and then, the images were geometric corrected; lastly, the optimum models
for chlorophyll-a concentration inversing were discussed for multi-temporal TM images. The water quality parameters
were measured on 21 sample points in Tai Lake, China monthly as the monitoring network. The chlorophyll-a
concentration inversing models were built using semi-empirical approach by the integrated use of multi-temporal remote
sensing data and in situ data. The spectrum character of chlorophyll was analyzed following other's studying. Then the
different composed bands and component modes such as TM4/TM3, (TM4-TM3)/(TM4+TM3), TM3*TM4/ln(TM1), etc.
were discussed for building the regression models. The inversing accuracy was evaluated by relatively error. The
optimum models were selected for each month by comparing the different models. It could be concluded that: The mode
of multi-temporal equations might be the same or similar for different month. But the coefficients were quite different;
the reflectance of TM3 and TM2 band were the most often used parameter for model building; the estimated accuracy
increased with raising chlorophyll-a concentration. For example, when the chlorophyll-a concentration was lower than
0.009mg/l, the estimated value was not so accuracy. But when the chlorophyll-a concentration raised to 0.05mg/l the
relatively errors for all samples were less than 30%.
Plant canopy reflectance is affected not only by the optical properties of canopy components, but also by canopy structure. In this paper, the radiative transfer model was used to simulate rice canopy bi-directional reflectance to determine its sensitivity to leaf area index (LAI) and inclination. In simulating canopy bi-directional reflectance over 400-940 nm, LAI was changed from 1 to 7 at an increment of 1; leaf inclination was changed from 50<sup>o</sup> to 85<sup>o</sup> at an interval of 5<sup>o</sup>. All other parameters in the model were measured in the field or deduced from references. It is found that with the rise in LAI, nadir reflectance decreases in visible light but increases in near infrared wavelengths. It tends to become stabilized when LAI is sufficiently large (e.g., >4). Decreasing with leaf inclination, canopy nadir reflectance becomes more sensitive to leaf inclination at a larger LAI. At 550nm and 670nm, bi-directional reflectance decreases with LAI regardless of view zenith. At 830nm, it is proportional to LAI over the view zenith angles of -85<sup>o</sup> - 40<sup>o</sup>. However, it is inversely related to LAI when it exceeds 3. Similar to nadir reflectance, bi-directional reflectance tends to become stabilized at a larger LAI at all view zeniths.
A modeling approach is used to assess the applicability of the derived equations which are capable to predict chlorophyll content of rice leaves at a given view direction. Two radiative transfer models, including PROSPECT model operated at leaf level and FCR model operated at canopy level, are used in the study. The study is consisted of three steps: (1) Simulation of bidirectional reflectance from canopy with different leaf chlorophyll contents, leaf-area-index (LAI) and under storey configurations; (2) Establishment of prediction relations of chlorophyll content by stepwise regression; and (3) Assessment of the applicability of these relations. The result shows that the accuracy of prediction is affected by different under storey configurations and, however, the accuracy tends to be greatly improved with increase of LAI.
The data for this study was collected from two-year (1999 and 2000) field experiments that based on different artificial nitrogen treatments. Linear, non-linear and stepwise multiple regression analysis were adopted for modeling The data in 1999 was utilized as training sample for modeling hyperspectra remote sensing estimation of rice aboveground fresh biomass, and the data in 2000 was evaluated and tested the models' predictive accuracy. Results of fitness analysis between hyperspectral variables and rice aboveground fresh biomass indicate that some hyperspectral characteristic variables and their combinations are closely correlated to aboveground biomass, such as red edge wavelength (λr),maximum reflectivity in green region, minimum reflectivity in red region, and the vegetation index based on the sum of first derivative spectral reflectance in blue region and that in red region. Determining the highest correlated wavebands and the best-fitting variables for raw spectra, first derivative spectra and hyperspectral characteristic variables through stepwise multiple regressions, and the results reveal that the relationship between the first derivative spectra and rice aboveground biomass is much clearer and simpler when compared with the rest. The best model for rice aboveground biomass estimation is based on the ration vegetation indices that calculated with the sum of the first derivative spectra reflectance in blue region and that in red region.
In this paper, we report some correlation analysis results between hyperspectral data in the spectral range of approximately 350nm-935nm and LAI of rice. Hyperspectral measurements were taken using an Analytical Spectral Devices (ASD) FieldSpec UV/VNIR Spectroradiometer at the experiment farm in 1999 and 2000.The potential of
hyperspectral data for estimating LAI was evaluated using univariate correlation method with different types of predictors: original and the first-order derivative spectra, vegetation index (VI) based, spectral position-based, area-based predictors. The 6 VIs were constructed from the green-peek and red-well spectra bands; spectral
position-based, predictors consisted of parameters extracted from the blue, yellow, and red edges, the green-peek and the red-well; area-based variables were calculated as the sum of the first derivative values at each of the three edges.
Results showed that for univariate correlation analysis, the better results were obtained for LAI. The best LAI was obtained with the area-based predictors in prediction models for LAI. In univariate correlation analysis, it seems that only wavelength at maximum value of 1st derivative within red edge (Wr), reflectance at green-peak and at red-well, and their VIs may be employed to predict LAI, and betterR<sup>2</sup> valued were obtained from the maximum first derivative spectra of blue edge (SDb) and red edge(SDr). In general, the results obtained from the accuracy assessment the best
predictors are area-based ones, the VIs of SDb and SDr.
Results from the correlation analysis showed that in the regions of the "three edges" for estimating LAI, sum of 1st derivative values within red edge was the most effective, sum of 1st derivative values within blue edge was the mere effective, sum of 1st derivative values within yellow edge was not effective.
To improve our understanding of photon transporting inside leaves, and hence improve the accuracy of yield estimating and growth monitoring of rice by remotely sensed data, we simulated rice leaf reflectance by PROSPECT model. The experiment, which were referred to as the late rice experiment, were conducted at Zhejiang University in 1999 and 2000 with one species of rice (which is called Xiushui 63); In 1999 the rice was planted normally, but in 2000 it was fertilized in three different levels (low, medium and high). Leaf spectrum (reflectance and transmittance), biochemical concentration such as chlorophyll, protein, cellulose, lignin and water content, and leaf area were measured during the experiment. By the PROSPECT model, we simulated leaf reflectance on four days’ data set in 1999 and one day’s data set of three fertilizations in 2000. The correlation coefficients between actual and simulated values are more than 0.995, the <i>RMSE</i> values are less than 0.0212. On the other hand, the model has been inversed to estimate chlorophyll concentration. Compared with actual value, the comparative errors are less than 10%.