Search

This work deals with multisensor data fusion to obtain landcover classification. The role of featurealevel fusion using the Dempster-hafer rule and that of data-level fusion in the MRF context is studied in this paper to obtain an optimally segmented image. Subsequently, segments are validated and classification ace curacy for the test data is evaluated. Two examples of data fusion of optical images and a synthetic aperture radar image are presented, each set having been acquired on different dates. Classification accuracies of the technique proposed are compared with those of some recent techniques in literature for the same image data. [ABSTRACT FROM AUTHOR]

To develop an operational methodology for estimating soil moisture and crop biophysical parameters and to generate a crop cover map, backscattering signatures of vegetation canopies are investigated using multitemporal Radarsat synthetic aperture radar (SAR) data over a predominantly cotton-growing area in India during low to peak crop growth stage. A simple parameterization of the water-cloud model with volumetric soil moisture content (mv) and leaf area index (LAI) is used to simulate the microwave backscattering coefficient (σ0), as it is found to be a good candidate for operational purposes as demonstrated by several workers in past. The influence of crop height (H), LAI, and mv on σ0 is investigated during peak crop growth stage. A linear relationship between LAI and crop height is derived semiempirically, and a linear zone is chosen for analysis during the peak crop-growing stage. Estimation of average volume fraction of leaves (&Vsline;l) and attenuation factor (L) by two different approaches is discussed: 1) using linear relationship between LAI versus crop height and 2) from the water-cloud model parameter (κ) estimation by iterative minimum least square error approach. It is observed that model-estimated parameters agree well with the measured values within an acceptable error limit. At tower soil moisture, mv ≅0.02 (cm³ · cm-3), the dynamic range of &cigma;0 is found to be about +5 dB for 0-70 cm of crop height but monotonously decreases to null at a transition point, having mv ≈ 0.38 (cm³ · cm-3). A positive correlation is found between backscattering coefficient and crop height fill this transition point but shows a negative correlation beyond that, signifying the predominant attenuation by vegetation over soil. [ABSTRACT FROM AUTHOR]