Digital Elevation Models (DEMs) can be generated using different techniques and data sources including levelling, photogrammetry, SAR interferometry, radargrammetry, and laser scanning. The different DEMs are affected by different systematic vertical and horizontal errors, as well as random noise. Moreover, each DEM has different spatial resolution and coverage. In general, the spatial resolution of a DEM is inversely proportional to the areal coverage. For example, airborne LiDAR DEMs have may have less than 50cm resolution and 1km swath length while DEMs generated from optical satellite data may have 1~10m resolution and 185km swath length, airborne SAR DEMs have less than 1m resolution and 16km swath, while the SRTM DEM has 30m resolution and 255km swath length. Obviously high resolution DEMs need many data to cover a large region for applications such as terrain monitoring. If all terrain areas had the same high spatial resolution, the resulting DEM would be huge. However, if only certain areas are of interest, such as CBD areas, industrial areas, or damaged areas, require high spatial resolution and other areas, such as forest, river, and mountainous terrain, have low resolution, the total data size and processing time is reduced. The authors propose a method that exploits the information contained in the area of overlap between different DEMs in order to reduce the total DEM data size and processing time by removing excess data in areas of reduced importance or interest, while at the same time reducing the vertical systematic errors.