Assessment of crop damage using space remote sensing and GIS

Abstract

This paper reports on a new Earth Observation (EO) research field concerning the potential use of space remote sensing for the assessment of crop damage at the field level. Digital field (cadastral) maps were used in order to overcome the problem of poor field boundary distinction (due to the current spatial resolution, small field size, a hilly landscape and the homogeneity of the area cover) and to estimate damage at field level. The relationship between crop damage estimations made by field observations and Normalized Difference Vegetation Index (NDVI) data was studied. By transforming the cadastral (field) map into a GRID format containing cells of one metre square (ArcView, Spatial Analyst), it was possible to determine the number of cells overlaying pixels or part of pixels only within the field area and the corresponding mean NDVI value. Various techniques, including Supervised Classification and Regression Models, were applied in order to study the correlation between NDVI values and those estimated by Hellenic Organization of Agricultural Insurance (HOAI) experts. The results appear to be very promising and the HOAI has decided to continue the research using the next generation high-resolution satellites.