We know that data comes in a variety of different formats and configurations, from a variety of sources, and each has its own importance. Our DataDoors solution is premised on the idea that users may want a variety of data types over any given area of interest. Our GAME application is also based on the idea that organizations may want to store many different types of data (including documents) in a single integrated application with a simple user interface.
For custom projects, we often work with clients who have concerns that any new data we create match up to their existing datasets. In such cases, we go the extra mile by treating the existing data as fixed and shifting the new data to match it near the edges.
Combining multiple sources
To get full coverage using satellite data, we recommend combining imagery from the various high- resolution satellites to form a complete, cloud-free mosaic. For example, an area may be partially covered by GeoEye-1 data, but when combined with WorldView-2 and QuickBird, the coverage is complete. When using multiple sources in a project, it is important to ensure that data is consistent spatially from one source to another, i.e. that there are no offsets from one to another. This may involve collecting common ground control points, orthorectifying higher-precision scenes to use as control, and/or generating a combined bundle adjustment. When sources are a similar type of data (for example, natural color imagery), we also color-balance the differing sources to each other, so the result looks like it came from one source rather than a patchwork.
Combining multiple resolutions
When imagery with vastly different resolutions are overlain, we can color balance vertically, providing the raw outputs needed to produce the appearance of seamlessly zooming in from the upper atmosphere down to the street level. One such product that has been very popular in both the weather broadcast and visual simulation industries is our MetroPak data. A MetroPak is a combination of low-resolution satellite imagery covering a broad area (typically centered on a city) with high-resolution imagery over the primary area of interest (typically the urban core). This can come from two or even three sources which differ widely in spectral characteristics. Using our proprietary PatchMatch technique, which leverages our proprietary differential superposition algorithm, we integrate these disparate datasets into one spectrally seamless stack of layers. Typical levels are 1m for the high- resolution data, 15m for a medium layer, and 100m for the overview level.
Combining different types of data
We can also combine two different types of source data into an exciting synergistic product. A good example would be a combination of an image mosaic with a hillshade of similar resolution. The resulting product still looks like imagery, but has a fun, three-dimensional look and feel.