Textural classification of B&W aerial photos for the forest classification

The Ministry of agriculture of the Czech Republic has defined a pilot project to summarize possible information that can be automatically evaluated from black and white aerial photos. This information should serve as input data into the large forest database or as signal data for forest state management organizations. These data were derived in traditional and modern ways. The traditional one used well-known principles of image processing as image distraction and thresholding. Modern tools were applied for other tasks using Fractal Net Evolution Approach commercially introduced by Baatz and Schäpe (1999) incorporated in commercial software eCognition for image segmentation and further classification where not only black and white aerial photos were used but also texture measures of these B&W aerial photos. The textural classification as another way used results of the detailed object oriented classification. The methodology was tested in another project defining the geodynamical model of land. The result of the project is a methodology to delineate forest areas, to distinguish deciduous and coniferous forest, to detect new deforestation and new large illegal dumpings and erosional rills from two different time level aerial photos. These tasks also include uninsured forest area detection. It means to determine six year-old forest (and younger).