The multispectral and multitemporal classification approach of AVHRR data on specific dates was studied for land cover mapping on a continental scale. Major conclusions are: decision keys must be developed that exploit both the use of multitemporal composites and multispectral data; the current classification distinguishes only main land cover types; due to mixed pixels the content of an AVHRR-derived land cover class can differ substantially over Europe. The main advantage of remote sensing over statistical surveys is its high spatial detail and consistent quality in land cover; the main disadvantage is the absence of specific classes.
|Place of Publication||Delft|
|Number of pages||103|
|Publication status||Published - 1996|