Detection of River Boundary Edges from Remotely Sensed Image

Detection of river boundary from remotely sensed imagery plays avitalrole in space-based river studies. Procuring of river attributes such as length, width, branching pattern, boundaries and temporal variation are very useful in several applications such as surface water supply, transport, distribution, and dynamics. Typically, field surveying is commonly used method to study of these river characteristics.It is commonly time-consuming, labor intensive and expensive method to gather river attributes in the field. In particular, it is unsafe and not feasible to measure rivers in certain environments such as ice sheets, tidal flats, and floodplains.Therefore, satellite imageriesof the earth surface are playing critical roles in river studies.The detection of linear and curvilinear components is a classic subject in image processing studies. In present study a method for the detection of river boundary is described. Image of Linear Imaging and Self Scanning Sensor (LISS-III) of Resourcesat-2 satellite is used. Method includes three basic steps i.e. mosaicking of different tiles of images, edge detection and connected component analysis.