Approach to Accuracy Assessment tor RS Image Classification Techniques

Abstract

Several techniques exist for remote sensing (RS) image classification, which includes supervised and unsupervised approaches. Classified maps are the main product of remote sensing image classification. Accuracy assessment of these classified maps is one of the foremost and important tasks of RS image classification technique. Without accuracy assessment the quality of map or output produced would be of lesser value to the end user. However, supervised and unsupervised techniques show different levels of accuracy after accuracy assessment was conducted. This paper describes a study that was carried out to perform supervised and unsupervised techniques on remote sensing data for land use/cover classification and to evaluate the accuracy result of both classification techniques. The study used IRS 1C LISS III satellite image consists of 26718 pixels, which covers Ralegaon Siddhi watershed in Ahmednagar district of Maharashtra state, India as a primary data and topographical map of SOI, cadastral map, and district statistical handbook containing land use/cover information as ancillary data. The land use/cover classes for the study area were classified into 5 themes namely, agricultur al land, shrubs, water body, wasteland and barren land. Ground verification was carried out to verify and assess the accuracy of classification. A several sample points with sufficient numbers of samples were collected based on Systematic Random sampling criteria. The comparative analysis based on the overall accuracy and Kappa statistics for the various classiers reflects the better performance of maximum likelihood classification technique.