Keywords: Growth estimation, Site-specific management, Hyperspectral reflectance spectrum, Normalized difference vegetation index, Satellite image of
3. Results and Discussion
Much research and progress have been
made in the areas of crop growth modeling and production estimation using the SIs incorporating SCs chosen from remote sensing data such as NDVI (Rouse et al. 1974; Yang and Chen 2004). This study measured the near-ground hyperspectral reflectance of rice canopy along plants development and identified the SCs in the red (RRED) and near-infrared (RNIR) regions to compute NDVINB. Plant samplings were made on days of spectral measurements to determine LAImeasured. The exponential relationship between LAImeasured and NDVINB
was developed (Figure 2) as that proposed by Yang and Chen (2004), and the regression equation derived was used as template to produce output values of LAIBB by using NDVIBB as inputs.
Change of LAI to NDVI is the most common relationship used to link growth parameter with the spectral data for growth monitoring and yield prediction (Aman et al., 1992; Price, 1992; Rajapakse et al., 2000). The exponential relationship implies that soil brightness produces large variation in NDVI at lower end of LAI when young plants are distributed sparingly.
Figure 2 Changes of the measured values of leaf
area index (LAImeasured) in response to the values of NDVINB from near-ground canopy hyperspectral reflectance data for rice plants (Oryza sativa L. cv. TNG 67) grown in the first and the second cropping seasons of 2006.On the other hand, leaves overlapping along vertical layers of canopy may change reflectance behavior with small NDVI variation at higher end of LAI when maturing plants toward the plateau of growth. Of the most importance, an input of NDVI is able to be transformed into a corresponding value of LAI with the regression equation (R2 = 0.781, P <
0.001), and hence, spectral variable (NDVI) obtained from remote sensing is interchangeable with growth parameter (LAI) as plants develop (Gong et al., 1995). However, many papers indicated that the relationship between LAI and NDVI is species dependent and may be affected by environmental factors (Baret and Guyot, 1991; Best and Harlan, 1985; Wanjura and Hatfield, 1987). Accordingly, the locational effects should be taken into consideration for a practical application and adjustments need to be made by considering the specific local conditions.
In contrast to NDVINB computed from narrowband reflectance of hyperspectral spectra, values of NDVIBB were calculated from broadband data of red and near-infrared regions of multispectral images taken by Formosat-2 satellite during the experimental periods of rice growth. By comparing the paired values of NDVIBB and NDVINB obtained from the same measuring days, it showed that NDVIBB values from Formosat-2 images correlated linearly (r = 0.793***) with NDVINB values from hyperpsectral reflectance (Figure 3). The positive correlation suggests that changes of NDVIBB follow the variation of NDVINB positively. Results also support the fact that spectral data of Formosat-2 are closely linked to spectral changes on the ground so that images taken by the satellite may be compiled and interpreted with ground truth.
Figure 3 The correlation between NDVIBB
from spectral images of FORMOSAT-2 satellite and NDVINB from near-ground canopy hyperspectral reflectance spectra for rice plants (Oryza sativa L. cv. TNG 67) grown in the first and the second cropping seasons of 2006.By the inputs of NDVIBB to the previously developed LAImeasured—NDVINB relationship
(Figure 2), values of LAIBB were derived as outputs. The correlation between paired values of LAIBB and LAImeasured was also linear statistically (r = 0.729***) (Figure 4). However, as in general NDVIBB values were lower than NDVINB values on the same measuring dates, the estimated value of LAIBB was smaller relative to that of the measured value of LAImeasured. A calibration should be made to compensate for this indigenous difference following the processing algorithms if an accurate estimation in needed.
Figure 4 The correlation between LAI
BB estimated from spectral images of Formosat-2 satellite and LAImeasured from plant samplings for rice plants (Oryza sativa L. cv. TNG 67) grown in the first and the second cropping seasons of 2006.4. Conclusions
As a result, the correlation of LAI and NDVI is high in rice crop so that LAI of rice crop can be reasonably estimated by employing the FORMOST-2 satellite high-temporal and high-spatial imagery. The approach proposed by Yang and Chen (2004) is proved applicable to the integrating of the FORMOSAT-2 observations with a comprehensive dataset containing canopy hyperspectral reflectance and plant samplings collected in the field for such purpose. However, there is still much can be done to improve the image processing algorithms for bettering NDVIBB acquirements and LAIBB estimation.
REFERENCE
Aman, A., H.P. Randriamanantena, A. Podaire, and R. Frouin. 1992. Upscale integration of normalized difference vegetation index: the problem of spatial heterogeneity. IEEE Transac. Geosci. Remote Sens. 30:326-338.
Baret, F., and G. Guyot. 1991. Potentials and limits of vegetation indices for LAI and APAR assessment. Remote Sens. Environ.
35:161-173.
Best, R.G., and J.C. Harlan. 1985. Spectral estimation of green leaf area index of oats.
Remote Sens. Environ. 17:27-36.
Bouman, B.A.M. 1995. Crop modeling and remote sensing for yield prediction. Neth. J.
Agric. Sci. 43:143-161.
Carter, G.A., and A.K. Knapp. 2001. Leaf optical properties in higher plants: Linking spectral characteristics to stress and chlorophyll concentration. Amer. J. Bot.
88:677–684.
Clevers, J.G.P.W. 1986. The derivation of a simplified reflectance model for the estimation of LAI. p.215-220. In M.C.J.
Damen, G. Sicco Smit and H. TH. Vers Tappe (eds.) Proceedings of Symposium on Remote Sensing for Resources Development and Environmental Management. August 1986, Balkema, The Netherlands.
Clevers, J.G.P.W. 1988. The derivation of a simplified reflectance model for the estimation of leaf area index. Remote Sens.
Environ. 25:53–69.
Coulter L.L., D.A. Stow, and S. Baer. 2003. A frame center matching technique for precise registration of multitemporal airborne frame imagery. IEEE Transac. Geosci.
Remote Sens. 41:2436-44.
Dai X, and S. Khorram. 1998. The effects of image misregistration on the accuracy of remotely sensed change detection. IEEE Transac. Geosci. Remote Sens. 36:1566-77.
Elvidge, C.D., and Z. Chen. 1995. Comparison of broad-band and narrow-band red and near-infrared vegetation indices. Remote Sens. Environ. 54:38-48.
GER-2600 User Manual. 1996. Version 1.1.
Geophysical and Environmental Research Corp., Millbrook, NY, USA.
Gilabert, M.A., S. Gandia, and J. Melia. 1996.
Analyses of spectral-biophysical relationships for a corn canopy. Remote Sens. Environ. 55:11-20.
Gong, P., P. Ruiliang, and J.R. Miller. 1995.
Coniferous forest leaf area index estimation along the Oregon transect using compact airborne spectrographic imager data.
Photogram. Eng. Remote Sens.
61:1107-1117.
Hall, A., D.W. Lamb, B. Holzapfel, and J. Louis.
2002. Optical remote sensing applications for viticulture—a review. Aust. J. Grape Wine Res. 8:36–47.
Khush, G.S. 2005. What it will take to feed 5 billion rice consumers in 2030. Plant Mol.
Biol. 59:1-6.
Knippling, E.B. 1970. Physical and physiological basis for differences in reflectance of visible and near-infrared radiation from vegetation. Remote Sens.
Environ. 1:155–159.
Lee, Y.-J., C.-M. Yang, K.-W. Chang, and Y.
Shen. 2008. A simple spectral index using reflectance of 735 nm to assess nitrogen status of rice canopy. Agron. J.
100:205–212.
Liu, C.-C. 2006. Processing of Formosat-2 daily revisit imagery for site surveillance. IEEE Transac. Geosci. Remote Sens.
44:3206-3214.
Liu, C.-C., J.-G. Liu, C.-W. Lin, A.-M. Wu, S.-H.
Liu, and C.-L. Shieh. 2007. Image processing of Formosat-2 data for monitoring South Asia tsunami. Intl. J.
Remote Sens. 28:3093-3111.
Liu, C.-C., C.-L. Shieh, C.-A. Wu, and M.-L.
Shieh. 2008. Change detection of gravel mining on riverbeds from the multi-temporal and high-spatial-resolution Formosat-2 imagery. River Res. Appl.
(accepted).
Liu, C.-C., and P.-L. Chen. 2009. Automatic extraction of ground control regions and orthorectification of Formosat-2 imagery.
(submitted).
Ma, B.L., M.J. Morrision, and L.M. Dwyer.
1996. Canopy light reflectance and field greenness to assess nitrogen fertilization and yield of maize. Agron. J. 88:915-920.
Pierce, F.J., P. Nowak, and P.C. Roberts. 1999.
Aspects of Precision Agriculture. p.1-84. In D.L. Sparks (ed.) Advances in Agronomy, Vol. 67. Academic Press, San Diego, CA, USA.
Price, J.C. 1992. Estimating vegetation amount from visible and near infrared reflectacnes.
Remote Sens. Environ. 41:29-34.
Rajapakse, R.M.S.S., N.K. Tripathi, and K.
Honda. 2000. Modelling tea (Camellia (L)
O. Kuntze) yield using satellite derived LAI,
landuse and meteorological data. p.1-6. In Proceedings of the 21st Asian Conference on Remote Sensing. Vol.1. December 4-8, 2000. Taipei, Taiwan ROC.Rouse, J.W., R.H. Haas, J.A. Schell, D.W.
Deering, and J.C. Harlan. 1974. Monitoring the vernal advancement and retrogradation (greenwave effect) of natural vegetation.
NASA/GSFC Type III report. Greenbelt, MD, USA. 371pp.
SAS Institute. 1998. SAS/STAT user's guide.
Version 8.1. SAS Inst., Cary, NC, USA.
Tucker, C.J., J.H. Elgin, Jr., J.E. McMurtrey, III.
and C.J. Fan. 1979. Monitoring corn and soybean crop development with hand-held radiometer spectral data. Remote Sens.
Environ. 8:237-248.
Wang, Y.-W. 2007. Integrating Formosat-2 High-Temporal And High-Spatial Imagery With Field Data To Monitor Growth And Estimate Yield Of Rice Crop. M.S. thesis, National Cheng Kung University, Tainan, Taiwan ROC.
Wanjura, D. F., and J. L. Hatfield. 1987.
Sensitivity of spectral vegetation indices to crop biomass. Trans. ASAE 30:810-816.
Wiegand, C.L., A.J. Richardson, D.F. Escobar, and A.H. Gerbermann. 1991. Vegetation indices in crop assessment. Remote Sens Environ 35:105-119.
Yang, C.-M., and R.-K. Chen. 2004. Modeling rice growth using hyperspectral reflectance data. Crop Sci 44:1283-1290.
Yang, C.-M., C.-H. Cheng, and R.-K. Chen.
2007a. Changes in spectral characteristics of rice canopy infested by leaf folder and brown planthopper. Crop Sci 47:329-335.