Automatic Change Detection from Time Series High Resolution Data

Abstract

Remote sensing techniques are vital especially for the study of land use /land cover, because of availability of suitable sensors operating at various resolutions and imaging scales. Disasters are a specific case for sudden change in land covers. Rapid response to disasters is necessary for coordination and mitigation. For example, to find changes along pipeline, within the right of usage is necessary since that damage can cause threat to life in the surrounding areas. Prevalent techniques for effective monitoring of change are visual interpretation, change vector analysis (CVA) transformation and artificial neural networks. For fully automated techniques, the important parameters to be considered for processing are time of imagery, resolution, look angles, radiometry and sensor types. The methodology proposed details incorporation of region growing methods and classification techniques. To overcome the disadvantages of manual trial and error procedures, automatic technique for the analysis of difference image is proposed. The main focus is change detection on single band panchromatic image atdifferent time series obtained from CARTOSAT1 / 2 sensors.

2 Figures and Tables

Cite this paper

@inproceedings{Poornima2014AutomaticCD, title={Automatic Change Detection from Time Series High Resolution Data}, author={U. S. Poornima and Reedhi Shukla and Ravi Kumar and Venugopal Rao and Raja Shekhar and Sri S. Sridhar}, year={2014} }