| A COMPARATIVE STUDY OF SUPPORT VECTOR MACHINE AND MAXIMUM LIKELIHOOD CLASSIFICATION TO EXTRACT LAND COVER AND LAND USE OF LAHORE DISTRICT, PUNJAB, PAKISTAN |
| Paper ID : 1084-SMPR |
| Authors: |
|
Fatima Mushtaq1, Khalid Mahmood *2, Muhammad Hamid Ch.3, Rahat Tufail4 1Center for Geographical Information System, Punjab University, Lahore 2Remote Sensing and GIS group, Department of Space Science, University of the Punjab, Lahore, Pakistan. 3Center for GIS, University of the Punjab, Lahore, Pakistan 4College of Earth and Environmental Science, University of the Punjab, Lahore, Pakistan |
| Abstract: |
| Mapping of Land Use and Land Cover (LULC) is very important because it directly influences environment in enormous ways and has become a vital variable of global change. In the last few decades, accurate thematic classification of LULC using remote sensing and GIS techniques has been emerged as a viable solution. In this regard various image classification techniques have been developed to minimize the error in mapping. This study focuses on the comparison of two of the supervised classification techniques for Sentinel-2 data, in order to select the more suitable method to extract land cover information of the Lahore district, Punjab, Pakistan. The selected classification methods are Maximum Likelihood Classifier (MLC) that based on neighbourhood function and Support Vector Machine (SVM) that based on optimal hyper-plane function. For this optimization, four land cover classes have been selected i.e. built up area, planted/cultivated land, open/bare land and water bodies. SVM is a non-parametric approach and relies on optimization of parameters while MLC is a parametric approach and relies on likelihood concept. Field based training samples have been collected and prepared through a survey of the study area at four spatial levels comprises of 100, 200, 300, and 400 polygons. Accuracy of results for each of the classifier has been assessed through error matrix and kappa statistics. Comparison of both classification techniques shows that SVM performs better than MLC. Overall accuracies of SVM and MLC are 95.20 % and 88.80 % whereas their kappa co-efficient are 0.9348 and 0.8462 respectively. |
| Keywords: |
| Land Use/ Land Cover (LULC), Support Vector Machine (SVM), Maximum Likelihood Classification (MLC), Accuracy Assessment, Kappa statistics |
| Status : Conditional Accept (Poster) |
