Open AccessLand use Land Cover Classification of Burhner River Watershed Using Remote Sensing and GIS Technique
Author(s) -
J. N. K. Rao,
S. K. Sharma,
M. K. Awasthi,
S. K. Pyasi,
S. K. Pandey
Publication year - 2022
Publication title -
international journal of environment and climate change
Language(s) - English
Resource type - Journals
ISSN - 2581-8627
DOI - 10.9734/ijecc/2022/v12i730707
Subject(s) - land cover , watershed , land use , remote sensing , cohen's kappa , ground truth , agricultural land , satellite imagery , geography , environmental science , hydrology (agriculture) , cartography , agriculture , computer science , artificial intelligence , geology , machine learning , civil engineering , geotechnical engineering , archaeology , engineering
Information on Land Use Land Cover (LULC) pattern is very important for developing management strategies for land use planning. Remotely sensed satellite data has proved to be an unmatched source of information that can deliver LULC information with good accuracy especially in regions where acquiring LULC information through intensive ground surveys seems to be impractical. The present study aims at performing LULC classification of Burhner river watershed situated in Mandla, Balaghat and Dindori districts of Madhya Pradesh, India by adopting unsupervised classification. It also discusses merits and demerits of using unsupervised classification for a high resolution satellite image along with key factors responsible for using it. Sentinel-2B satellite imagery with spatial resolution of 10 m was used in the present investigation. A total number of 6 LULC classes were identified in the study area namely agricultural land, fallow/open land, forest, habitation, wasteland and waterbodies using ERDAS IMAGINE® 2011. Accuracy assessment of the LULC classified image from reference (ground truth) data using error matrix revealed an overall accuracy of 95.72% with kappa coefficient of 0.94. Furthermore, the error matrix also aided in computing classified image producer’s and user’s accuracy which were under acceptable limits. The LULC statistics of study area indicated that highest area is covered by forest (53.01%), followed by fallow/open land (24%), agricultural land (19.44%) and least area is covered by waterbodies (1.38%) and habitation (0.19%). A major portion of study area under fallow/open land category pointed that underutilized land resource potential exist in watershed. Such land can be further utilized for crop production and plantation purposes in order to maximize output from the available natural resources in a sustainable manner.