Field Data Collection Program for Accuracy Assessment of Flood Mapping of Microwave Remote Sensing

GIC staff member Dr. Jagath Rajapakse has conducted a field data collection program for accuracy assessment of flood mapping in Colombo, Sri Lanka during 13-17 May 2016. The main focus of the program was to assess the accuracy of satellite products provided by Sentinel Asia (SA) during the recent flooding in Sri Lanka.
Sentinel Asia (SA) is a framework which mandates to provide satellite data at a time of a disaster as well as coordinating support combining satellite data providers, data analyses nodes including disaster management organizations in the region. During the past six years (2010-2015), 63% of the SA activations have been for floods which was the most frequent disaster in 2015.
However, the accuracy of flood mapping is still an issue as water detection algorithm is hampered by different land use and land cover types. The accuracy assessment of flood maps produced from SAR (Synthetic Aperture Radar) is important for post disaster activities and mitigation planning. Heavy flooding in Sri Lanka, occurred in mid May 2016, was selected for the accuracy assessment and field data collection was conducted with the local support from the Survey Departments of Sri Lanka. The main objectives were;

  • To evaluate the accuracy of flood maps generated from various SAR bands based on land cover types
  • To evaluate the water detection accuracy based on polarizations available in different SAR sensors
  • To determine the best band and polarity and thresholding for water detection on different land-cover types