Implications of interpreting tropical dry forest succession after radiometric correction of Chris/ Proba images

Abstract

In the present paper, we evaluate the effect of cast shadows in the classification of three successional stages of tropical forest in Mexico using hyperspectral and multiangular Chris/Proba images. A simple algorithm based on the cosine of the angle of solar incidence on the terrain is applied to correct the effect of topography on Chris/Proba's reflectances. Previous to the correction of topographic effects, Chris/Proba images were atmospherically corrected in BEAM software. Vegetation maps of the study site were generated using non-parametric decision trees, defining four mam classes: late, intermediate and early stages of succession within a tropical dry forest, and riparian forests. By the comparison of resultant vegetation maps before and after the correction of topography in Chris/Proba's spectral data, we found that late stage of succession and riparian forests are overestimated in non-corrected images while intermediate and early stages of succession are underestimated. Errors in classification are more important in extreme Chris/Proba's angles of observation. Thus, the radiometric correction of topography is necessary for an accurate classification of succession in tropical dry forests.