This study aimed to correlate logging intensity with canopy opening using medium-resolution satellite images (SPOT-type) on a sample of 15 blocks totaling more than 3300 ha in two French Guianan forests with different reliefs. The maps obtained show the cumulative impacts on the canopy over the entire logging period (2008-2010). The percentages of canopy openness and the unit areas per logged tree were examined with logging statistics and GPS mappings of felled stems for 4 sub-samples. At the block scale, a 20% canopy opening after logging was observed for a logging intensity of 3.5 trees ha_1. The average canopy opening size per harvested tree was 601 m2, which included felling gaps and openings resulting from the construction of roads and trails needed to remove the timber. The coefficients of variation of these variables were low. General linear models (glm) were created to test the relationship between canopy damage, logging intensity and relief at local scales (1 and 4 ha) and at the management unit scale (i.e. 200-300 ha). The model created at management-unit level (block level) did not detect any significant effect of the variables on the percentage of canopy openness. However, a significant and concordant relationship was demonstrated at local levels between relief, logging intensity, their interaction on the one hand and the percentage of canopy openness on the other. A maximum local harvesting intensity of 8 trees per hectare on plateaus makes it possible to obtain an opening of less than 33% at an 85% probability threshold. In contrast, the same canopy openness threshold is obtained at only 5 trees per hectare on hillier terrain, which suggests that it would be advantageous to adjust silviculture recommendations according to forest landscape. The average openness of 20% observed in this sample, considered in relation to the cutting cycle in effect in French Guiana (65 years), implies a complete renewal of forest stands after 325 years, assuming multifunct