Application of Object-oriented Classification with Hierarchical Multi-Scale Segmentation for Information Extraction in Nonoc Nickel Mine, the Philippines

At present, high spatial resolution remote sensing images have been widely applied in the ground objects classification. However, the information extraction of typical opencast mining areas by using high spatial resolution remote sensing images is less studied. The open-pit Nonoc nickel mine is one of the largest laterites nickel ore in Philippines. In this paper, based on the data resource of high spatial resolution remote sensing images, we used the method of object-oriented classification with hierarchical multi-scale segmentation to extract ground object information in the Nonoc nickel mining areas. The qualitatively and quantitatively relative analysis between single scale and hierarchical multi-scale identification results shows that hierarchical multi-scale segmentation has better effect and the highest precise, and the overall accuracy and Kappa coefficient are 92.73% and 0.9024 respectively. Consequently the hierarchical multi-scale segmentation method is more suitable to be applied to the information extraction of open-pit laterites nickel mining areas.