Object-Based Classification of IKONOS Data for Endemic Torreya Mapping

Torreya (T. grandis ‘Merrillii’) is a rare cash crop tree endemic to Kuaiji mountains of southern China. The present research demonstrated the use of spatial information in object-based classification for improving the identification of Torreya from IKONOS images. Pixel-based and object-based classifications on a combination of spectral and textural features were implemented and their accuracies were compared. The object-based classification proved more accurate than the pixel-based classification. The overall classification accuracy and Kappa coefficients were 75.7% and 0.676 for the pixel-based classification; 81.7% and 0.746 for the object-based classification. The overall results demonstrate that object-based classification of very high resolution satellite images using additional features is a practical tool for vegetation mapping in the study area.