Introductory Digital Image Processing: A Remote Sensing Perspective
Introductory Digital Image Processing: A Remote Sensing Perspective
A fast semi-inverse approach to detect and remove the haze from a single image
ACCV'10 Proceedings of the 10th Asian conference on Computer vision - Volume Part II
| Hi-index | 0.00 |
In the framework of a land cover study of a reclamation area, a high-resolution IKONOS scene characterized by an extremely non-homogenous haze cover was processed in order to reduce haze and calculate some vegetation indices (VIs). Haze removal was performed by means of a technique based on the haze optimized transform (HOT). As HOT is also sensitive to components other than haze, with respect to the original procedure we introduced a second step in which the haze component estimated from HOT was masked both for water bodies (as in the original approach) and urban features (responsible for high HOT values). Interpolation was finally performed in order to reconstruct haze in such areas. Dark-object subtraction (DOS) was then performed on the dehazed scene in order to estimate approximate soil reflectances necessary to calculate VIs. Their calculation for original and dehazed data illustrates the beneficial effects of the haze removal process and shows that, in the case of severe haze cover, even so-called atmospherically resistant VIs require a preliminary haze removal procedure.