Vegetation canopy structure plays an important role in the partitioning of incident solar radiation, photosynthesis, transpiration, and other scalar fluxes. The vertical foliage distribution of the plant canopy is represented by leaf area density (LAD), which is defined as the one-sided leaf area per unit volume. Airborne light detection and ranging (LiDAR) offers the possibility to characterize the 3-D variation of LAD over space, which still remains a challenge to estimate. Moreover, the low density of point cloud data generally offered by airborne LiDAR may be insufficient for accurate LAD estimation in dense overlapping forest canopies. We develop a method for the estimation of the LAD profile using a combination of airborne LiDAR and hyperspectral data using a feature-based data fusion approach. After identifying vegetation species using hyperspectral data, point cloud LiDAR data is used in a “tree-shaped” voxel approach to characterize the LAD of trees in a riparian forest setting. We also propose a set of relationships on simple geometry of overlap for the construction of tree shaped voxels. In a forest setting with overlapping canopies, the results indicate that the tree-shaped voxels are better able to attribute the LAD to the upper and middle parts of the overall canopy as well as individual tall and short trees compared with traditional cylindrical voxels.