Extending Time-Series of Satellite Images by Radiometric Intercalibration

Abstract

In recent years, the necessity to assess the state and development of large environmental units has gained increasing attention. In this respect, remote sensing based approaches can provide a synoptic, repetitive and coherent perspective over large areas. The derivation of concise conclusions largely depends on the quantitative consistency of the data employed; hence, a completely standardised analysis of multi-temporal time series is mandatory. In the context of the required radiometric rectification, sensor calibration is an essential component, the accuracy of which largely determines the overall result. While for newer sensor systems extensive information on the sensor’s sensitivity exists, older systems often lack the corresponding information. A methodology has been developed to radiometrically intercalibrate different sensor systems based on the precondition of synoptic image acquisition, which is for example provided by the Landsat-TM and -MSS sensor systems hosted on Landsat 5. It supports a full radiometric rectification of Landsat MSS data and enables their incorporation into satellite image time series, which can thus be significantly extended.