|
The key research task of this project is to test
the separability of irrigated crops under different conditions such as
scale and size of land units, complexity and heterogeneity of landscape
and land use, seasonal meteorological conditions - rain/no rain,
temperature and atmospheric conditions - clear/cloudy/hazy. Conventionally, in higher resolution image
analysis, the conditions within a pixel is assumed as homogenous. In
practice, The use high resolution images for global scale studies is
rather difficult and expensive. The coarse resolution imagery implies the need for
sub-pixel techniques to identify the proportion
of a specific land-cover that is likely to be contained within that pixel.
In most real-world cases, a MODIS pixel at 500m by 500m (25 ha) will not
contain 100% irrigated area. In areas with large scale contiguous surface
irrigation, there might be roads, field bunds, channels and houses in an
otherwise fully irrigated landscape. In more complex situations there may
be other land use types – forest, fallow, grassland and so on. Different
methods need to be used to relate the proportion of irrigated land in a
coarse pixel observable remote sensing information, such as
reflectance, temperature and backscatter values.
On going research topics related to the sated issues are given below:
| |
