Creating a Mosaic Dataset

Nowadays, there are many sources where we can get remotely sensed data: satellite imageries, aerial photos, digital elevation models, IFSAR, LiDAR, and more. Organizations with large collections of data sources may encounter difficulty in managing and making it accessible to all their users. 

Traditionally, we share our files physically that may create problems, especially with large datasets. To resolve this, we can use Esri's Mosaic Datasets. Mosaic dataset manages your data as a catalog collection where you can dynamically deliver a mosaicked image on-the-fly.

So, how do we create a mosaic dataset? First step is to check if your dataset needs to be preprocessed - build pyramids and calculate statistics. By building pyramids, you create a reduced-resolution versions of your datasets that will then enable faster access. Calculating statistics on the other hand ensures that your dataset is rendered properly. After that, you can now create your mosaic dataset.

Here are the typical steps:

1. Create a Geodatabase.
2. Create a mosaic dataset inside the Geodatabase that you have just created.
3. Add all your raster data.
4. Define additional metadata - e.g. sensor name, product type, processing template, filter, minimum and maximum stretch, band combination, pan-sharpening technique, orthorectification method and function templates.
5. After that, the raster data will be cataloged inside your Mosaic Dataset. You can now refine the footprints of your raster data and define NoData pixels as well.
6. Then, you can create seamlines. Seamlines are used to define the line along which the rasters in the mosaic dataset will be mosaicked.
7. After that, you can color balance your mosaic to create a seamless looking mosaicked image. This is necessary since your mosaic dataset is composed of many raster datasets that may have significant color differences.
8. Then, you can refine your mosaic dataset.
9. Lastly, you will need to generate overviews. Overviews, in a way, are like raster dataset pyramids.

If you haven't tried mosaic datasets, try creating one now! xx