Remote Sensing of Desert Cities in Developing Countries - Remote Sensing of Urban and Suburban Areas - page 248

In addition, the study focused on urbanization encroachment, population/demography,
and agricultural practices on the urban-rural fringe.

Satellite and Ancillary Data Acquisition : In order to perform the study, satellite
imagery was required at the appropriate period of time (year, season). Existing car-
tographic data (topographic maps, road network maps, boundaries) are also neces-
sary to be used in the process of registration and integration with satellite imagery
and/or to supplement the analysis with the necessary input in dates for which satel-
lite imagery could not be acquired. Ground observations and existing statistics could
also be used in visual interpretation and the validation of remote sensing results.

Land Use Mapping Using Satellite Imagery : Two multi-date satellite images were
used for each city. The adopted interpretation technique was based on visual inter-
pretation assisted by computer. The process consisted of displaying of the rectified
imagery on the screen and the digitizing of polygons, which represent different land
use categories, based on image characteristics such as tone, texture, color and pat-
tern that can be translated into land use attributes. The whole process was guided
by ground observations and local knowledge. These polygons were identified by
independent labels attached to the polygon centeroids. The ground truthing aimed
at localizing and characterizing land use categories using GPS instruments. The
result of the visual interpretation technique was a digital coverage layer, in which
polygons represented different land use categories. A full database is attached to
this coverage, in which every polygon is characterized by many attributes like, the
category number, area and perimeter of polygons.

Data Analysis and Interpretation Using GIS : We utilized the pair-wise comparison
of land use maps, as well as their statistical inventories, in order to detect changes
both in terms of area and geographic location of change. Urban and sub-urban areas
were identified and assessed, in order to produce a map between two periods
depicting changes in only urban and sub-urban areas and to investigate the impact
of urbanization on other land uses and vise-versa. An “evolution matrix” was estab-
lished for the two periods to provide all possible paths of change.

13.4.2 The Case Study of Morocco

Context : The main objective of this study was the assessment of urbanization
impact on agricultural lands in the three cities of Ksar El Kébir, Béni Mellal, and
Khémiset, during the last 3 decades (1970s, 1980s and 1990s). This multi-temporal
study has been conducted in order to inform the urban planning scheme (Shéma
Directeur d'Aménagement Urbain). Each study site covered a square of 30 km by
30 km centered on each of the cities. Two panchromatic Spot-HRV images with 10 m
spatial resolution were used to cover the last 2 decades. In addition, topographic
city maps were used to cover the first decade (1970s). Ground observations were
used to complete the process of validation and control of land use mapping.