Automatic classification of digital orthophotomap of agricultural landscape
2005
Szocsová, I.,Soil Science and Conservation Research Institute, Bratislava (Slovak Republic)
Classification as one of the image processing methods brings simplification into obtaining of a big amount of information contained in satellite or aerial imagery. The aim of this study is to show possibility of extracting agricultural parcels from orthophotomap using object oriented approach. The inspiration along with challenge to use this method in agriculture was given by the existence of LPIS (Land Parcel Identification System) in Slovak Republic as a part of IACS (Integrated Administrative and Control System). The process of the object oriented classification involves two steps, multiresolutional segmentation and classification. As each image analysis deals with structures of a certain spatial scale, the average image objects size must be free, adaptable to the scale of interest. This is achieved by a general segmentation algorithm based on homogeneity definitions in combination with local and global optimization techniques. The scale parameter is used to control the average image object size. Different homogeneity criteria for image objects are based on spectral and/or spatial information. Object features are used by means of fuzzy logic to build class descriptions. The result of the classification is a network of classified image objects with concrete attributes, concrete relations to each other and concrete relations to the classes in the class hierarchy. Classification was evaluated by comparison of the results with the LPIS layer and vectorized orthophoto.
显示更多 [+] 显示较少 [-]