Research Methodology - Historical Land Use/Land Cover Classification Using Remote Sensing - page 136

Fig. 5.53 The followed concept to classify the major winter crops based on both the previous
general classes classification and the generated spatial distribution of the irrigated areas
crops of interest (wheat, barley, and sugar beet), in addition to fallow, which is
classified in both seasons.
Trees and shrubs could be classified from either winter remotely sensed data or
from data acquired in summer. This information was then combined and a mask
constructed that included all the winter-data-based classified LULC-areas which
could be planted during the summer (fallow, wheat, barley, and other crops). The
other classified areas which were almost impossible to be changed during the
summer of the same year (e.g., trees, permanent crops, etc.) were excluded.
Finally, the before built mask was applied to the summer-related remote sensing
data (Fig. 5.54 ) for classification of the summer crops of interest (cotton, and
corn), plus fallow. Before the completion of this task, it was necessary to perform
an accuracy assessment for all the produced classification results.
The other method that was adopted to classify the irrigation areas (especially
the state projects) and the various permanent, winter, and summer agriculture-
categories, was to integrate the remotely sensed data with the non-remotely sensed
data (e.g., statistical records and detailed schemes), especially for the TM-May and
August-1987 data. This is explained in the next section.
The most important point involved in these projects was the engineering
organization and division of each project into several farms with names and known
geographical sites. The cultivated fields were large enough to be easily distin-
guished by available remotely sensed data. In addition, the geographical distri-
bution of those fields had well known coordinates and detailed charts, and