Identification and classification of landslides type using spatial and spectral features procedures by object-oriented approach (Nasirabad to Sattarkhan dam of Ahar Chay)

Abstract

Diagnosis and classification of landslides is a critical need in the risk analysis before and after the disaster. And primarily through land surveying or traditional interpretation of images was done. In this paper to identify and classify types of object-oriented approach landslide has been paid. The data used in this study consisted of false color images obtained from satellite data Resourcesat-1 with spatial resolution of 5.8 meters and digital elevation models with 2.5-meter resolution satellite image of 10 meters of Cartosat-1 was used. This method was used for the North West basin and then used without further reforms in the eastern part of the basin. A total of three sliding using this method accurately identified 71.11% and 91.4% classification accuracy has been detected. In this way, the landslide early detection of high accuracy and speed, hence has great potential to assist in risk analysis, disaster management and decision making process after the earthquake or heavy rainfall, can be used related entities, including crisis management headquarters, natural resources and watershed institutions.

Keywords

• landslide
• object-oriented approach
• identification
• segmentation
• irs satellite