LSMA in Precise Forest Classification
Nikolov, Hristo; Borisova, Denitsa
Space Research and Technology Institute - Bulgarian Academy of Sciences (SRTI-BAS), BULGARIA

During the recent years many field observations have been made close to two open pit mines (operating and non-operating). The most serious damages caused by mining industry are made to the near forest areas. In the process of subsequent monitoring and land reclamation the precise allocation of the forest types in the close areas is essential. The correct determination of suitable forest species for replanting may rely on historical data from satellite missions prior to opening the mine. In the framework of this study we put the availability of multispectral data with different spatial and temporal resolution from airborne sensors (satellite and aerial) concerning the region of interest. The LSMA was applied to assess the proportion between low and high vegetation and between coniferous and deciduous trees. The data we have used in this research are from different sources namely raster imagery from different years over the same area (starting from 1978) and vector data from CORINE projects. Also some in-situ multispectral data from the region under study was collected by the authors using the Thematically Oriented Multichannel Spectrometer, designed in SRTI -BAS and data from spectral libraries for the same tree plants. As an outcome from our studies we prepared maps of the region close to the mining processing plant providing good estimation of the vegetated surroundings.

Figure 1. Using well proven discrimination features for both possible tree species in LSMA