Open AccessClassification of Landsat-8 Imagery Based On Pca And Ndvi Methods
Author(s) -
M Venkata Dasu,
Dr P V N Reddy,
Dr S Chandra Mohan Reddy
Publication year - 2019
Publication title -
international journal of innovative technology and exploring engineering
Language(s) - English
Resource type - Journals
ISSN - 2278-3075
DOI - 10.35940/ijitee.j9843.0881019
Subject(s) - normalized difference vegetation index , remote sensing , cohen's kappa , satellite , vegetation (pathology) , pixel , contextual image classification , computer science , principal component analysis , satellite imagery , spectral bands , artificial intelligence , pattern recognition (psychology) , image (mathematics) , geography , geology , machine learning , medicine , oceanography , pathology , climate change , aerospace engineering , engineering
Remote sensing is an important issue in satellite image classification. In developing a significant sustainable system in agriculture farming, the major concern for remote sensing applications is the crop classification mechanism. The other important application in remote sensing is urban classification which gives the information about houses, roads, buildings, vegetation etc. A superior indicator for the presence of vegetation can be computed from the vegetation indices of a satellite image. This indicator supports in describing the health of vegetation through the image attributes like greenness and density. The other parameter in detecting objects or region of interest is an image is the texture. A satellite image contains spectral information and can be represented by more spectral bands and classification is very tough task. Generally, Classification of individual pixels in satellite images is based on the spectral information. In this research paper Principle component analysis and combination of PCA and NDVI classification methods are applied on Landsat-8 images. These images are acquired from USGS. The performance of these methods is compared in statistical parameters such as Kappa coefficient, overall accuracy, user’s accuracy, precision accuracy and F1 accuracy. In this work existing method is PCA and proposed method is PCA+NDVI. Experimental results shows that the proposed method has better statistical values compared to existing method.