Pixel size ranges used

LegacyLegacy:

ArcGIS 10 is the last release of the stand-alone ArcGIS Image Server product. The image service definition (.ISDef) has been replaced by an improved geodatabase data model—the mosaic dataset—which can be published as an image service using the ArcGIS Server Image extension.

The pixel size ranges are used to determine what rasters are processed to create an image for the client. You can view the pixel sizes in the service table. The LoPS and HiPS columns define the actual range of pixel sizes that the raster dataset contains. For example, a raster dataset that contains pixels of 1 meter and a pyramid with pixel sizes 2, 4, 8, and 16 meters would have a LoPS of 1 and a HiPS of 16. Below is an example of a service table. The Generic:Tiff rows represent the original raster datasets and have the lowest pixel size ranges specified. There are three levels of service overviews—the bottom ServiceOverview has the highest value for the pixel size ranges and the coarsest resolution; the resolution improves as you move up. You can see that by specifying these ranges, the service overviews will be used to generate the client's image that has a coarser resolution than the original dataset (when the client is zoomed out viewing an overview of the image service), thus saving processing time. To learn more about service overviews, see Deriving images for image service definition optimization.

example of pixel sizes in the service table

The pixel size values are populated when the raster datasets are added to the image service definition and are always defined in meters irrespective of the spatial reference system used. These pixel sizes are only used to identify the range of pixel sizes and are not used when georeferencing the rasters; thus, the accuracy of the values has little importance.

The MinPS and MaxPS columns define the range of pixel size requests for which the raster dataset should be used. These values are populated when you check the Compute Pixel Size Ranges option while building an image service definition. You can also manually edit the values in the service table. The MinPS and MaxPS values are computed based on the overview and overlap between the images.

You can take advantage of the pixel size ranges when using multiple resolutions of raster datasets for a single image service definition. For example, at a countrywide level, you could use Landsat satellite images with approximately 30-meter pixel size to display when the client requests a particular scale and area of interest. However, as a client zooms in, requesting a larger scale, the image service can provide a SPOT image with a 5-meter resolution, and as the user zooms closer, they will eventually see aerial orthophotography, which has a much smaller pixel size (higher resolution).

You could also take advantage of the pixel ranges when generating overviews. Overviews can be generated for use when the client is requesting data at particular scales. Therefore, an overview can be created for all the images that will display when the user's area of interest covers the entire mosaicked area, and other overviews can be generated for different scale ranges—with the scale range being specified using pixel size ranges.

The LoPS and HiPS values are determined when the raster data is added to the image service definition and should not be changed. The ranges, MinPS and MaxPS values, are computed when an image service definition is built and can be manually changed if required. You can also override the MinPS and MaxPS values in the Output Definition section on the Service Properties dialog box, where the specified values override the raster dataset-specific parameters and stop imagery from being returned outside the specified ranges for the whole image service.

There is a direct relationship between the pixel size and scale. Both can be calculated using the following:

Scale = Pixel Size x 96/0.0254
Pixel Size = Scale x 0.0254/96

The pixel size is used in ArcGIS Image Server, instead of scale, because a pixel size is an inherent value of a raster dataset.

Related Topics


4/19/2011