A Comparative Study of Tasselled Cap Transformation of DMC and ETM+ Images and their Application in Forest Classification

In this paper, an attempt has been made to derive the tasselled cap coefficients manually following the rigorous mathematical computations for the three-band Disaster Monitoring Constellation (DMC) data. Considering ten DMC images of same season, it has been found that the first two transformation components, namely brightness and greenness can capture on an average 94.44 % of the three band variance of the DMC image. The coefficients of TC transformation equations for DMC image were derived considering the above mentioned ten DMC images. For the application purpose the TC images were applied to classify the forest types of the New Forest area of UK. For the comparison purpose, the tasselled cap images generated after the automated transformation of the Landsat 7 ETM+ image of the same area were considered to classify the forest types. A series of statistical intra-image and inter-image comparisons were conducted to find out the intra and inter-relationships between the brightness and greenness images obtained from the two different sensors. Finally, the accuracies of both the classified images were assessed based on the field collected GPS data. It was estimated that the overall accuracies of the classified DMC and ETM+ images were 77.44 % and 81.43 % respectively.