Especially in shaded areas a visual interpretation of 
vegetation cover is hardly possible. Here the vegetation 
index separates sealed from unsealed areas in detail, 
although the index is slightly raised in shadow zones. 
The vegetation index NDVI can be transformed for each 
pixel with a linear equation using the extreme values P,ealed 
(100 % sealing, NDVi eq) and Paes (0 % sealing, 
NDV\nsealea) IN quantitative sealing degree values. 
Another analysing method is to identify visually a threshold 
operator and to use it to transform the vegetation index 
data set into a binary picture. 
The advantage of this simple and practical method faces 
the following methodological disadvantages: 
incorrect sealing degree values for water areas: 
Water has a high absorption in red and near infrared 
spectral range. Thus the calculated vegetation index of 
water is very low and, after transformation, its sealing 
values are very high which is wrong. Because of this 
mistake it is necessary to mask water areas after having 
carried out a land-use classification or to exclude these 
areas by digitizing and masking them before classification. 
incorrect sealing degree values of unsealed areas Without 
vegetation: 
Vegetationless and unsealed surfaces cause a low 
vegetation index and consequently a high sealing degree 
value. These areas are therefore wrongly calculated, 
Number and expansion of such areas are very limited in 
urban areas. They appear as construction areas, disposal 
sites, quarries, fields and very intensively used allotments, 
They can be identified visually and need to be masked, 
4.3 Multitemporal Thermal Infrared Data as an Instrument 
to identify Soil Sealing 
Thermal infrared data give information about surface 
temperatures. They are mostly dependent on surface 
material, but also influenced by relief, daylight and wing 
exposition, as well as air humidity. The mean surface 
temperatures of different materials are determined for each 
of the three flight dates. 
Table 5 shows the temperature values and the calculated 
cooling values from late evening to early morning. 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
Surface material Midday temperature in °C | Evening temperature in °C | Morning temperature in °C Cooling rate in °C 
(ca. 13.15) (ca. 21.30) (ca. 4.45) (evening to morning) 
Mean Stand. Dev. Mean Stand. Dev. Mean Stand. Dev. Mean Stand. Dev. 
Tarmac road 38,4 3,3 25,0 1,3 16,6 0,8 8,5 08 
Paved road 33,7 2,1 22,9 1,0 15.5 0,9 6,1 0,9 
Concrete road 35,0 1,5 25,1 1,1 16,4 0,8 8,9 0,9 
Tile roof 38,0 7,0 18,7 1,9 10,1 157 8,7 1,6 
Bituminous roof 40,9 5,8 21,9 1,8 12,9 2,2 8,5 1,1 
Bare soil 38,1 4,3 18,6 2,5 11,9 2,3 6,8 0,8 
Uncultivated field 36,7 2,9 15,7 27 11,0 1,6 5,0 1,6 
Cultivated field 27,2 3,5 14,9 1,3 10,1 0,8 4,8 13 
Grass 42,1 24 17,2 1,2 10,8 1,0 6,4 0,6 
Meadow 32,4 5,4 16,0 1,8 9,5 17 6,1 A0. | 
Tree tops 23,6 2,0 20,9 0,5 12,1 1,2 84 BÉ 
Water 18,6 0,3 21,4 0,2 19,6 0,2 1,9 02 | 
  
  
  
  
  
Table 5: Temperature Values of Different Surface Cover Materials 
International Archives of Photogrammetry and 
  
  
  
  
  
490 
Remote Sensing. Vol. XXXI, Part B7. Vienna 1996