干旱区生态环境知识资源中心

A methodology for calibrating multispectral IRS Wide-Field Sensors (WiFS), which cover typically about 700 km swath is suggested. Two available bands B3 and B4 of IRS-1C and -1D WIFS and respective bands from four of IRS-P6 Advanced WiFS (AWiFS) are employed. Data products acquired over the Thar Desert within 5 days (near-synchronous condition) were utilized. Linear regression model is assumed valid to obtain the calibration coefficients (CCs), after compensating for intrinsic variations of these bands due to spectral response functions and solar zenith angles. It is found that IRS-1C and -1D WIFS cameras show a strong linearity relationship, as evident from their data samples with goodness of fit of better than 0.99. The application of the CCs may hence provide the desired integrated information when their data products are used in tandem. Analysis of the IRS-P6 AWIFS and IRS-1D WIFS image pairs, however, shows a large variability in their output, especially for the Band B4 datasets. The CCs derived for these sensors’ pair needs to be used cautiously, Variation in their relative spectral responses demands further investigation, probably with inclusion of the in-situ measurements to account for variations of the target reflectance and the atmosphere while attempting cross-calibration analysis.