Abstract: To solve the problem of large-scale urban true orthophoto's generation, the paper proposes an occlusion detection based on the overall projection of digital building model (DBM). Using the characteristics that the surface of DBM storage with triangular facets and raster inside projection of plane graphics do not shelter, we use triangular face as a unit to orthographic projection of the entire buildings to get polygon of the roof. Make perspective projection of the entire buildings to get the polygon of the building on the image, then get the polygon of the whole building on traditional orthophoto based on DEM projection iteration. We can get the occlusion areas of the building by subtracting two polygons. Finally, we get the true orthophoto after repairing the occlusion areas with the best image. The experiment shows that the method can detect the occlusion areas quickly and accurately and provide a prediction to generate high-quality true orthophoto.