No 1 Optimisation, Modelling, Analysis, Simulation and Scheduling (OptiMAS) Research Group
Remote sensing data recorded from passive satellite system tend to be degraded by attenuation of solar radiation due to haze. Haze is capable of modifying the spectral and statistical properties of remote sensing data and consequently causes problem in data analysis and interpretation. Haze needs to be removed or reduced in order to restore the quality of the data. This study aims to restore the hazy data using proposed haze removal technique and evaluate its performance by means of spectral
... eans of spectral and statistical methods. In this study, initially, haze radiances due to radiation attenuation are removed by making use of pseudo invariant features (PIFs) selected among reflective objects within the study area. Spatial filters are subsequently used to remove the remaining noise causes by haze variability. The performance of hazy data restoration was evaluated using Support Vector Machine (SVM) classification. It is revealed that the technique is able to improve the classification accuracy to the acceptable levels for data with moderate visibilities and restored the spectral and statistical properties of the data and shows an increase in overall classification accuracy from 51.63% to 82.62%.