EFFECT OF SPECTRAL CHANGE WIDTH IN RADIOMETRIC NORMALIZATION OF MULTITEMPORAL SATELLITE IMAGERY METHODES EFFICIENCY

摘要

Multitemporal satellite optical images of the same terrain are very applicable in monitoring and quantifying large scale land cover change over time. These images are confounded in terms of radiometric consistency due to differences in sensor calibration parameters, illumination, geometric condition and variation in atmospheric effects. To analyze these images, it is necessary to omit mentioned differences. Relative radiometric normalization is used to prepare these images for stated applications. In this paper, different relative radiometric normalization methods are compared with each other. Images of Landsat-5 thematic mapper (TM) and Landsat-7 enhanced thematic mapper plus (ETM+) captured from Iran are used for this comparison. The results of various techniques have been evaluated with both visual inspection and statistical analysis i.e. Root Mean Square Error (RMSE) and Universal Quality Indicator (UQI) value between each pair of analogous band. Obtained results show intensity value of spectral changes affects quality of normalization. In images which have small area with intensified spectral changes, MS, HM and MM methods produce better results; while in images with intensified changes covering large area, PIF method generates the best results. Also among mentioned methods, MS method is the most stable with respect to changes.