Respond Projects

Research Areas

Cloud avoidance scheduling

Payload programming makes the optimum use of satellite resources to satisfy User requirements. The various capabilities of the IRS satellites and its resources call for a meticulous planning. Payload programming is successful only when it results in acquisition which cater to user requirements in terms of data quality, correct area of coverage (targeting accuracy), timeliness (within the period of interest) and cloud free acquisition. Cloud cover is one of the major problems in the acquisition of optical satellite remote sensing data and has a negative impact on the efficiency of data scheduling. The necessary global cloud information (on a daily / hourly basis) derived from meteorological satellites has to be incorporated in the planning system to improve the planning efficiency.

Algorithms for time series analysis of satellite data

National Remote Sensing Centre has been archiving all remote sensing data right from IRS 1A mission of ISRO till date. The datasets has all the characteristics that may constitute a time series and hence is a good resource for long term studies related to climate change and change detection etc. One of the possible applications of a time series data is to build models for forecasting. Statistical Methods such as ARIMA and GARCH are popular in predicting time series but they are far from satisfactory in terms of precision. Recent developments in machine learning algorithms and its use in regression and classification problems have paved a way for their use in forecasting time series. We invite proposals that incorporate the state-of-the-art algorithms for remote sensing applications, characterization of sensors and reconstruction of data.