An effective hybrid classification approach using tasseled cap transformation (TCT) for improving classification of land use/land cover (LU/LC) in semi-arid region: a case study of Morva-Hadaf watershed, Gujarat, India

Abstract

In recent decades, the semi-arid Morva-Hadaf watershed from Gujarat, India, is experiencing rapid land use/ land cover (LU/LC) changes due to watershed management measures. The accurate LU/LC mapping and change detection of these areas are crucial for impact assessment studies and also for policy making. Remote sensing (RS) is an efficient and cost-effective means of monitoring landscapes. One of the key challenges in the RS study is to improve the accuracy of classification, which is quite complicated in heterogeneous semi-arid regions because of spectral similarity among the different LU/LC features. To improve the classification accuracy, in the present study, the effectiveness of conventional supervised classification with only visual bands was enhanced using (1) supervised classification with only tasseled cap transformed (TCT) components and (2) hybrid classification with TCT components and ancillary topographical data of slope and aspect. The proposed hybrid classification approach has enhanced the classification accuracy significantly by 8.13 and 7.81 % for the year 1997 and 2011, respectively. The change detection using hybrid classification showed a significant increase in agricultural area from 593.62 to 713.01 km2 and simultaneous decrease in the area under scrub forest/land with scrub class from 252.85 to 154.43 km2. It has indicated an overall positive impact of watershed management measures in Morva-Hadaf watershed.