VisibilityIndexOperator.OutputResolutionFactor Property
Gets or sets an integer factor that modifies the output resolution.
Syntax
public AttributeVariable< System.Int64 > OutputResolutionFactor { get; set; }Property Value
Default: 1
An integer factor that modifies the output resolution.
Remarks
By setting the OutputResolutionFactor to n, you say that you desire output raster cells that are n times larger than the basic resolution in which the calculations are done. This basic resolution is often the finest raster resolution available from the data source for the ElevationInput, but it is possible to choose a coarser basic resolution for the calculations by increasing the DesiredInputResolution.
Whatever the basic resolution is, the OutputResolutionFactor provides a way to make the output resolution even coarser by the given factor. However, the implementation will round the factor down to a power of 2. In other words, a factor n = 1 gives the same output resolution as the basic resolution, n = 2 or 3 gives twice as large cells, n = 4, 5, 6 or 7 gives four times as large cells, etc.
Output rasters with lower resolution will have lower accuracy, but they need less memory and are faster to compute.
For example, let us assume that elevations are available in 10 meter resolution, with overviews of 20, 40, 80 and 160 meters. The default settings for DesiredInputResolution and OutputResolutionFactor would make both the basic resolution and the output resolution 10 meters, which may turn out to be too slow. So if you think you would be satisfied with an output resolution of 40 meters, there are several ways you can choose the values of the two properties:
Let DesiredInputResolution be 40.0 and let OutputResolutionFactor be 1.
Let DesiredInputResolution be 20.0 and let OutputResolutionFactor be 2.
Let DesiredInputResolution be 10.0 and let OutputResolutionFactor be 4.
Although the different ways will give the same output resolution, they will differ in speed and accuracy.
Letting OutputResolutionFactor = 4 in this example will be safer in terms of accuracy, since it approximates the result of first doing a calculation in a basic resolution of 10 meters and then resampling the result to 40 meters by a Max filter. The emulation of the resampling does not always find the true Max, but fairly often, and the resampled value will at least be between the Max and the Min. To be precise, the operator will use a quick rule of thumb to find the most promising 10 meter cell inside the 40 meter output cell, calculate its visibility index in the basic 10 meter resolution, and store the result in the 40 meter output cell. The most promising cell is the one with the highest ground elevation, excluding cells where the trees are higher than the sensor.
If you instead let DesiredInputResolution = 40.0, there is no such guarantee of accuracy relative to the 10 meter result, but the calculation can be faster. Of course, thanks to the better speed, you could choose to trade some of speed improvement for better accuracy by using a larger number of sample Rays.
In summary, it is non-trivial to find the best way of combining the two properties to get a desired output resolution and accuracy.
Platforms
Windows, Linux, Android