Information extraction of mining-land use based on automatic threshold classification of decision tree

This paper presents automatic threshold classification of decision tree. Taking Lu'an mine area as an example, the land use information from 1993 and 2000 year's TM images are extracted. The classification results reach relevant accuracy requirements, from which 1993-2000 conversion matrix of land use is attained. That will help to analyze the causes of the land-use changes so as to provide a strong basis for mining-land reclamation planning.