Classification of landsat 8 satellite data using unsupervised methods

Abstract

The information from band extraction and calculation of indices which are used for classification imagery of Landsat 8 satellite data using unsupervised methods were studied. The visible and Near Infrared (NIR) bands of Landsat 8 satellite were used to derive Normalized Different Vegetation Index (NDVI) image. The Normalized Difference Water Index (NDWI) is a satellite-derived index from the NIR and Short Wave Infrared (SWIR) bands. Vegetation, non-vegetation, and water features classes were then analyzed by classification experiment of three unsupervised methods: ISODATA, K-means, and fuzzy c-means with guidance of ground truth information of the study area. The accuracy of the classified image is then assessed using a confusion matrix where classification accuracy and kappa coefficient are computed. The result shows that unsupervised methods classification is able to classify the Landsat 8 satellite data with FCM got a high accuracy compared to another two methods.