Determination of land cover change by multi-temporal radar imagery

Vietnam is an Asian country with a hot and humid tropical climate, monitoring the land cover with optical satellite images is usually used. However, because of the cloud cover, the optical satellite image is not good. To overcome these effects, Radar image is a useful option, which the radar pulses can penetrate through the clouds, regardless of day or night. In this paper we mention the use of multi temporal C band satellite images to monitor land cover changes of an area in the Mekong Delta, Vietnam. The image used is Sentinel-1, a free image provided by the European aerospace agency (ESA). The set of 3 images is March 2015, March 2016 and February 2017. Double polarized images (VV, VH) were subsetting for VH polarization to determine the land cover changes. With 3 time of images which were composited, then using Random Forest (RF) classification method to detect the change of some land cover types. Evaluate the accuracy for determining by RF classification achieved over 80%. © 2020 40th Asian Conference on Remote Sensing, ACRS 2019: "Progress of Remote Sensing Technology for Smart Future". All rights reserved.