LAND USE LAND COVER (LULC) CLASSIFICATION USING FUSION OF MULTI-SENSOR OPTICAL and MICROWAVE DATA FROM SENTINEL

Hits: 25
Research areas:
Year:
2017
Type of Publication:
Article
Keywords:
Multi-sensor fusion,LULC,Sentinel Mission, SAR, Ehlers Fusion
Authors:
Vivek Kumar, Ashutosh Bhardwaj A.L. Haldar
Abstract:
Land Use land Cover (LULC) mapping from satellite imagery is of great important since it allows to analyze terrain features and is also useful for monitor temporal changes (change detection) like dynamics of water resource, forest cover or urban environment economically. Optical satellite sensors usually detect reflection from features of earth in the visible and infrared part of the electromagnetic spectrum. In contrast, Synthetic Aperture Radar (SAR) has ability to penetrate cloud and also independence of requirement of daylight. These advantages make SAR remote sensing technique different and attractive data source for land-use land-cover mapping. In this study the objective is the mapping of Land use land cover using both Synthetic Aperture Radar (SAR) & Optical remotely sensed datasets through fusion techniques.Sentinel-1A and sentinel-2A mission datasets are used here for parts of Uttar Pradesh .The classification of fused data has been done with help of HPF (High pass Filter) resolution merge and Ehlers fusion methods. In addition both the sentinel imageries from the sentinel mission has proved to be the best contemporary possible open source images which suit the study. The fusion of SAR data (sentinel 1A) and multispectral image (sentinel 2A) has improved the separability of classes and hence the accuracy is improved from 76.17% in sentinel 2A image to 82.42% and 82.81% in the HPF and Ehlers fused images respectively. Thus, the study depicted that the resultant fused image from multisensory data has better quality for classification.
Full text: 72.pdf [Bibtex]
You are here: Home ACRS ACRS Overview Proceedings land-use-land-cover-lulc-classification-using-fusion-of-multi-sensor-optical-and-microwave-data-from-sentinel