To manage your scanned maps, you will use one source mosaic dataset. The workflow tools automate many of the best practices outlined here. These specifications are provided mainly for reference.
Source mosaic datasets
Your collection of scanned maps should be stored as a source mosaic dataset, optimized to render the data seamlessly and without collars.
The Create Source Mosaics tool streamlines the creation of source mosaic datasets using the Table raster type. This tool uses the MDCS script, with parameters set using the appropriate XML configuration file. The following MDCS configuration files are included with the workflow tools:
| Configuration file name | What does the config file do? |
|---|---|
| scannedmaps_table.xml | Defines the parameters that are used by MDCS to create a source mosaic dataset of scanned maps. |
When creating my mosaic dataset, what parameters should I use?
When managing scanned maps, there are three aspects of a mosaic dataset that are important to keep in mind.
- Querying. Users will likely want to query the scanned map service you publish, so the option to allow queries should be checked on when you publish your service, and field names users are likely to query should be clear and understandable. However, the services you create are likely going to be large. To conserve transmission bandwidth, it's useful to turn off fields that aren't likely to be queried by the user.
- Displaying metadata. Displaying metadata provides the user with useful context. Identify the most relevant fields and allow transmission of those fields.
- Time awareness. If you want to use a time slider with your mosaic dataset to view your maps, you need to make your service time aware. Time aware services should have a date field that contains valid dates and time for each record (usually available from the map's metadata). To make a service time aware, you have to set up the time field at the same time you set up the service. For Scanned Maps, the year on the map is often the best option.
Additionally, when creating a new source mosaic dataset, use the following parameters:
| Parameter | Guidelines |
|---|---|
| Spatial reference system | WGS_1984_Web_Mercator_Auxiliary_Sphere |
| Number of bands | Default |
| Bit depth | Default |
| Time Enabled | Yes, if Date field exists |
What should the mosaic dataset properties be?
Recommended properties for a scanned map mosaic dataset are below. If you are using the workflow tools, these parameters will be set automatically.
| Parameter | Recommended setting |
|---|---|
| max_num_per_mosaic | 50 |
| rows_maximum_imagesize | 4000 |
| columns_maximum_imagesize | 4000 |
| default_compression_type | JPEG |
| JPEG_quality | 85 |
| resampling_type | BILINEAR |
| clip_to_footprints | CLIP |
| clip_to_boundary | NOT_CLIP |
| footprints_may_contain_nodata | FOOTPRINTS_DO_NOT_CONTAIN_NODATA |
| allowed_mensuration_capabilities | BASIC |
| default_mensuration_capabilities | BASIC |
| allowed_mosaic_methods | Center; NorthWest; Nadir; Viewpoint; LockRaster; ByAttribute; Seamline; None |
| default_mosaic_method | Seamline |
| sorting_order | Ascending |
| mosaic_operator | FIRST |
| blend_width | 0 |
| max_num_per_mosaic | 50 |
| cell_size_tolerance | 100 |
| cell_size | # |
| metadata_level | BASIC |
| geographic_transform | # |
| max_num_of_download_items | 200 |
| max_num_of_records_returned | 1000 |
Additionally, if your mosaic dataset is time-enabled, set the following default properties by editing the configuration file.
| Property | Set manually or by script? | Requirements |
|---|---|---|
| Time: Use Time (Yes/No) | Manually | If Date field is available |
| Time: Start Time Field | Script | Will need to edit the configuration file. |
| Time: End Time Field | Script | Will use the Date_On_Map field. |
| Time Value Format | Script | Formatted as YYYY |
| Time Interval Value | Depends on data | |
| Time Interval Units | Depends on data |
What parameters should I use when adding rasters?
If a feature table is available, you should add rasters to your mosaic dataset using the Table raster type. However, scanned maps in TIF or JPEG format can also be added to your mosaic dataset using the 'Raster Dataset' raster type.
The advantage of using the Table raster type is that the neatline footprint and any relevant metadata available in the feature table can be imported. Defining a field called "Raster" containing the full path to the TIFF file in the feature table ensures that the feature table is compatible with the Table raster type.
The workflow tool will automate the process of adding rasters, but for more information see documentation on adding rasters to mosaic datasets.
Use the following parameters when adding scanned maps.
| Parameter | Guidelines |
|---|---|
| Raster type | Table |
| Update Cell Size Ranges | NO_CELL_SIZES (See below for how to calculate cell size ranges) |
| Update Overviews | NO_OVERVIEWS |
| Duplicate Items | EXCLUDE_DUPLICATES |
| Statistics | NO |
| Pyramids | NO |
| Force Spatial Ref. | YES |
| Update Boundary | YES |
How should I populate metadata?
The most important fields are added to the mosaic dataset attribute table automatically by the workflow tools, and populated based on the metadata of the source scanned maps.
How should I deal with NoData values and footprints in my imagery?
Typically, there will not be any NoData values in a scanned map. The footprint for each image should show the entire image, excluding the collar or borders of the map.
Whats the best way to calculate sOrder and cell size ranges?
These fields are calculated using the Calculate Value operation. If you're using the workflow tools, the parameters can be set in the Calculate Value node in the configuration file.
sOrder is calculated using the following formula.
sOrder = [Map_Scale] - [Date_On_Map]
Cell size ranges should be calculated using a formula rather than reading the values from the TIFF file. Each scanned map will not have the same cell size—the cell size would be determined based on the actual physical size and resolution of the scan.
Due to the variance in cell sizes, the cell size tolerance factor should be set to 100. The MinPS and MaxPS values are calculated separately using the following formulas:
MinsPS = 0
MaxPS = [Map_Scale] / 2000
Do I need to create seamlines?
Yes. Seamlines should be used to clip out the collars of your scanned maps (so they should contain only the edited footprint showing the map, not the collar or the borders).
To accomplish this, the default mosaic method is set to seamlines. These seamlines are copied from the footprint, and the mosaic dataset property of 'clip_to_footprint' is set to false. This property, when used in conjunction with lock raster, will allow the user to see the mosaicked scanned maps in their entirety, with the map collars hidden but preserved.
Properties used while building seamlines are below. All other values should be set to default.
| Property | Setting |
|---|---|
| computation_method | COPY_FOOTPRINT |
| blend_width | 0 |
| blend_type | # |
Do I need to create overviews?
Optional. The best option is to use existing maps at smaller scales. Once the data is properly ingested and the appropriate mosaic method selected, overviews may be generated.
If you are using the workflow tools, best practices for building overviews have been defined in the configuration file. The only input required from the user is the pixel size. However, since map scale is widely used for scanned maps, it's a good idea to determine the map scale at which you wish to generate the overviews. If you're familiar with the data, this can be done as the data is being ingested. Otherwise, this can be done at a later stage, after you've examined the map scale values in the mosaic dataset.
See the next section to learn more about publishing your managed scanned maps.