True Orthoimage Production for Forested Areas from Large-Scale Aerial Photographs

production is relatively mature. These techniques are also used Abstract to produce digital orthophotos from high-resolution SPOT sat- An is a fundamental information source in forestry. ellite images (Chen and Lee, 1993). Orthophoto production is usually based on digital elevation Large-scale generation challenges the existing models (DEM); however, the elevations of 3D objects such as algorithms. Scale and view angle are crucial in determining buildings and trees, which are visible in large-scale photos, algorithm complexity. Landsat images generally do not contain are not included in DEMs. Recently initiated orthophoto significant relief displacement due to the coarse scale and the generation has been dedicated to urban scenes, removing the relatively narrow field of view (FOV). Such displacement how- displacement caused by both the terrain and buildings. This ever must be taken into account for aerial images at finer scales paper reports an effort on reducing occlusion and distortion and/or wider FOV. Generally, small- to medium-scale aerial caused by trees in true orthoimage generation for forested orthoimages are generated with DEMs. Consequently, only the areas. Canopy surface models (CSM) depicting both the terrain distortion caused by terrain relief is removed, but not that and tree surfaces were used in forest orthoimage production. caused by object (e.g., buildings and trees) relief. As photo Z-buffer techniques were incorporated into indirect