Using quickbird images fore annual monitoring of forestland

Abstract

The annual assessment of forestland is generally carried out at an administrative level by recording the occurred events. In Italy the forest areas subjected to harvesting are censused by the Italian National Statistical Institute: however various Authors point out some shortcomings in the adopted census procedure. Very high resolution remotely sensed images allow to test multitemporal and multispectral objective analyses and to evaluate the events occurred within the forestland. The aim of this note is to present the results of an experimentation of a multitemporal normalized index, to identify cutting areas in coppice woods performed in a year, using very high resolution satellite images. The index has been calculated for a couple of multitemporal QuickBird satellite images for a test area of 2215 ha in northern Latium (Italy). The identified cutting areas have been compared with the official administrative data, recorded according to the conventional procedures. The potential of the proposed procedure has been assessed by object oriented classification techniques. The results highlight the effectiveness of the proposed method for the quantification of the events occurred in the test area.