Classifying Remote Sensing (RS) imagery for reliable and accurate Land Use/Land Cover (LU/LC) change information still remains a challenge in heterogeneous and arid landscapes due to spectrally similar LU/LC features. The aim of the present study is to extract reliable LU/LC information using ancillary data and change detection between 2001 and 2011 for Arjuni watershed from highly arid state Gujarat, India using RS and Geographic Information System (GIS). The Maximum Likelihood Classifier (MLC) was first applied to IRS LISS-III imagery of 2001 and 2011 and classified as: water body, forest, agricultural land, scrub forest/Prosopis, barren land, settlement/built-up land, river sand and quarry. Further, the study employed an innovative methodological framework of ancillary data (viz., texture imagery, Normalized Difference Water Index-NDWI and drainage network) for post-classification corrections. It has significantly improved overall classification accuracies from 67.84% to 82.75% and 71.93% to 87.43% for 2001 and 2011, respectively. The change detection study showed an increase in agricultural land, forest, water body classes and decrease in scrub forest/Prosopis and river sand classes over the period of ten years.
Thakkar, A. K., Desai, V. R., Patel, A., & Potdar, M. B. (2017). Post-classification corrections in improving the classification of Land Use/Land Cover of arid region using RS and GIS: The case of Arjuni watershed, Gujarat, India. Egyptian Journal of Remote Sensing and Space Science, 20(1), 79–89. https://doi.org/10.1016/j.ejrs.2016.11.006