Résumé
In this study, waveforms acquired by the Geoscience Laser Altimeter System (GLAS) were used to analyze the difference between GLAS and SRTM elevations in the aim of discriminating the five forest landscape types (LTs) in French Guiana (Table 1). First, two differences were calculated: (1) penetration depth defined as GLAS highest minus SRTM elevations, and (2) GLAS centroid minus SRTM elevations. The behaviour of these two variables was studied for the different LTs as a function of GLAS canopy height and SRTM roughness index. Next, the classification potential of the five forest landscape types (LTs) using the coupling of GLAS and SRTM was assessed using the Random Forest algorithm. This classification was conducted using the penetration depth, difference between GLAS centroid and SRTM elevations, GLAS canopy height (Hc) (difference between the ground peak and signal start) and SRTM roughness index (R). Finally, using the changes occurring to the GLAS signal in different seasons, the potential of LT discrimination using these changes was studied.