International Archives of Photogrammetry and Remote Sensing, Vol. 32, Part 7-4-3 W6, Valladolid, Spain, 3-4 June, 1999 
Settlement planning: 
The settlement area has been more expanded 
compared to the planned area 
Existing settlement area: 
The demarcation of the current existing settlement 
area is possible. 
Settlement layer: 
settlement area (regional plan) 
planned settlement area (regional plan) 
as classified settlement structure 
(Image data: © ANTRIX, SIE, Euromap Neustrelitz) 
Scale 
2000 o 
Meters 
Fig. 1. Classified settlement structures compared to the regional plan. 
planning uses scales 1 : 50,000 and 1 : 100,000. Therefore, a 
certain generalization of the settlement surfaces is necessary. 
Moreover, very large buildings with strongly reflecting roof 
materials (metal, glass) are not detected by the morphological 
analysis at all or only partially. Likewise problematic are large 
traffic areas, partly outside the settlement areas, dirt roads and 
field edges, which are classified as settlements due to their 
strong reflection gradients. These structures can be eliminated 
however by overlaying of the class ‘agriculture’ from the 
original LISS data (23 m). 
A remaining class, uncertain sealed areas, can be extracted from 
the results of the multispectral classification, by masking out the 
classes that were determined with high accuracy (forest, water, 
sand, fallow land and construction areas). Within settlement 
areas, this class includes all built-up and sealed surfaces, which 
is desirable for the creation of a settlement mask. To avoid false 
classifications outside of the settlements, particularly within the 
agricultural areas, a ‘matrix analysis’ was used. This procedure 
produces a thematic layer that contains a separate class for every 
coincidence of classes in the two layers containing the 
multispectral and morphological partial results, i.e. a calculation 
of an ‘overlap index’ between the multispectral and 
morphological partial results was performed, followed by a 
buffering. The final results (see Figure 1) were very promising 
with a roughly estimated accuracy of 90-95% for the 
identification of settlement areas (the exact assessment is still to 
be done). 
5. URBAN STRUCTURE-TYPE CLASSIFICATION 
Extended discussions with municipal surveyors on the 
suitability of IRS colour composites led to the following 
conclusions. Mapping of urban structure types is important in 
urban ecology, since structure types relate to significant 
ecological features, e.g. they can be used to estimate the degree 
of surface sealing. Urban structure types can be detected with 
excellent clarity using a 1 :25,000 scale image. A visual 
structure-type mapping using the IRS-1C colour composite and 
a digital building-block map (city map containing building 
blocks of all structure types, separated by roads), is carried out 
for the City of Dresden, following the classification scheme 
listed in Table 2. 
The following assessments, regarding the usefulness of the IRS 
colour composite, were made by urban environmental 
protection authorities: An estimation of urban greenery is 
eminently feasible. The vitality of roadside vegetation can be 
determined, and parklands clearly subdivided into meadows and 
woodlands. The principal classes of urban biotope mapping can 
164