CART-Based Rare Habitat Information Extraction For Landsat ETM+ Image

In the paper, The DT classifier adopted was CART(Classification and Regression Trees) to obtain a habitat of interest to Paeonia sinjiangensis, existing as fragments on the southern coteau of the Altay mountain located Xinjiang autonomous region of China. In the paper, decision tree classifier allows for the integration of remotely sensed data with other sources of georeferenced information such as land use data, spatial texture, and digital elevation models (DEMs) to obtain greater classification accuracy. TM reflectance data acquired in 2001 were required to completely cover Altay mountain area. Several ancillary datasets were used as inputs into the decision tree classification. These datasets include land use, city boundaries, vegetation types and digital elevation models. Logical decision rules, discovered from samples through CART integrating spectral textural and the spatial distribution character, then are used with the above various datasets to assign class values to each pixel. Finally, In contrast with CART, a standard maximum likelihood decision rule implemented by a discriminant analysis. Results on study of Paeonia sinjiangensis ground extraction from ETM imagery show that the classification results of CART was significantly better than that of Common DT classification. And this organizational methodology for classification is feasible and reliable if taking advantage of ancillary data and image analyst for classification.