Landscapes Forest and Global Change - ESA - Escola Superior ...

M. Akbarzadeh & E. Kouhgardi 2010. Mapping and monitoring land cover and land –use changing using RS and GIS
332
First, land sat Thematic Mapper (TM) and Enhanced Thematic Mapper (ETM+) images
acquired on 20 July 2002 and 19 July 1987, respectively. Land sat 4 and 5 carry the TM sensors.
Second, digital topographic maps digitized from hard copy topographic maps with scale of
1:50,000 were made of IRANIAN surveying center and used mainly for geometric correction of
the satellite data and for some ground truth information.
Finally, ground information was collected between 1997 until 2002for the purpose of supervised
classification and classification accuracy assessment.
2.3 Geometric correction
Accurate per- pixel registration of multi- temporal remote sensing data is essential for change
detection since the potential exists for registration errors to be interpreted as land cover and
land- use change, leading to an overestimation of actual Change(stow, 1999).
Change detection analysis is performed on a pixel – by- pixel basis, therefore a misregistration
greater than one pixel will provide an anomalous result of that pixel. To overcome this problem,
the root mean – square error (RMSE) between any two dates should not exceed 0.5 pixels (Lune
tta& Elvidge 1998).
In this study geometric correction was carried out using ground control points from topographic
maps with scale of 1:5000 produced in 1986 by Iranian Army Surveying Center (IASC).To
geocode the image of 2002, then this image was used to register the image of 1987.
The RMSE between the two images was less than 0.4 pixel which is acceptable. The RMES
could be defined as the deviations between GCP and GP locations as predicted by the fitted –
polynomial and their actual locations.
The goal of image enhancement is to improve the visual interpretability of an image by
increasing the apparent distinction between the features.
The process of visually interpreting digitally enhanced imagery attempts to optimize the
complementary abilities of the human mind and the computer.The mind is excellent at
interpreting spatial attributes on an image and is capable of identifying obscure or subtle
features (Lillesand & Kiefer, 1994).
Contrast stretching was applied on the two images and two false color composites (FCC) were
produced.
These FCC were visually interpreted using on screen digitizing in order to delineate land cover
classes that could be easily interpreted such as village.
Some classes were spectrally confused and could not be separated well by supervised
classification and hence visual interpretation was required to separate them.
2.4 Image classification
Land cover classes are typically mapped from digital remotely sensed data through the process
of a supervised digital image classification (Campbell, 1987), (Zobeyri & Daleki, 1988). The
over all objective of the image classification procedure is to automatically categorize all pixels
in an image and to land cover classes or themes (Lillesand & Kiefer, 1994).
The maximum likelihood classifier quantitatively evaluates both the variance and covariance of
the category spectral response patterns when classifying an unknown pixel so that it is
considered to be one of the most accurate classifiers since it is based on statistical parameters.
Supervised classification was done using ground chek points and digital topographic maps of
the study area.
The area was classified into seven main classes: High density forest, low density forest,
ranglands, agriculture, bare land, river, and village.
Description of theses land cover classes are presented in Table 1.
Forest Landscapes and Global Change-New Frontiers in Management, Conservation and Restoration. Proceedings of the IUFRO Landscape Ecology
Working Group International Conference, September 21-27, 2010, Bragança, Portugal. J.C. Azevedo, M. Feliciano, J. Castro & M.A. Pinto (eds.)
2010, Instituto Politécnico de Bragança, Bragança, Portugal.