Using Coherence Change-Detection With Sentine1-1 For Natural And Man-Made Disaster Monitoring In Urban Areas

This study shows the value of globally available remote sensing imagery in disaster monitoring and management.The disasters have been divided into man-made/anthropogenic disasters(war)and natural disasters(earthquake,hurricane and fire).These different disasters have dissimilar effects on urban areas due to their varying nature.The study incorporates globally available Sentinel-1 data into disaster monitoring by using the Coherence Change Detection(CCD)method.Coherence loss is a sign of change and a disaster event is signified by coherence loss.The author observed that for each case study used,different land use classes respond differently to coherence loss especially after disaster occurrence.This is essential when selecting the appropriate technique or data to use when monitoring a particular disaster so as to minimize the social,economic,environmental etc.damages that may occur as result of the disaster.Therefore,assessing to what extent can Sentinel-1 data be used for natural and anthropogenic disaster monitoring,is the aim of this research.This research considers the scientific questions:How can changes from anthropogenic and natural disasters be detected using the CCD?How do different urban land use classes respond to coherence loss after a disaster has occurred?Which disaster can best be monitored using the CCD technique?To answer these questions,the author selected Sentinel-1 imagery as it has global coverage and is freely available.Furthermore,the archive of global images allowed the author to acquire a time series of a stack of images over a longer time.The data also has the advantage of weather independence,all-day and all-night availability and is very useful in detecting urban built up areas since it uses a sensor that operates in the microwave portion of the electromagnetic spectrum.Ultimately,the coherence images provide useful information to detect changes that have resulted from disaster occurrence.The coherence data was aggregated into land use class polygons derived from Landsat 5 and Landsat 8 imagery so as to observe the response of each land use class to coherence loss.Classifying coherence loss is also valuable in determining whether Sentinel-1 imagery through the CCD technique is useful in disaster monitoring by showing how urban classes are affected by decorrelation,which may also be useful in using coherence maps for classification in future studies.Further,the coherence data was aggregated to street blocks to calculate the standard deviation of each street-block.Averaging coherence images of urban street blocks and calculating the standard deviation can be useful tools in investigating and monitoring urban areas affected by disasters at street block level.Street-blocks are considered as relevant units of measurement in our study areas.Street-blocks,are defined as the smallest unit or ground parcel in an urban area surrounded by a road.We selected these to be elementary units for data aggregation.Furthermore,these street block are objects that relatable to the real world.Statistics is derived about the change over time and after disaster occurrence,through averaged coherence values and calculating standard deviation overtime,is also an essential part of detecting changes.