The DIRSIG LIDAR/LADAR model produces focal plane arriving photon counts within user-defined time gate. The time gate description includes the minimum and maximum time after each pulse that the photon counts will be recorded, and a delta time resolution within that window. This data cube then needs to be processed by a detection model and processed by a geolocation algorithm to translate these photon counts into a 3D point cloud that is the common product from such a system.
The DIRSIG model is distributed with a very simple program that will convert the time-gated photon count data cubes generated by the "lidar" capture method into XYZ point clouds. This program is called fxyz_convertor, but it is intended to be used as an example for a more elaborate detection and geolocation processing program.
With a LIDAR system, the photon counts are:
A combination of the passive (Sun, Moon, sky, etc.) and active (laser) photons.
Focal plane reaching instead of aperture reaching.
When modeling a 0 watt laser during the day and a passive system with the same general specifications (focal length, pixel size, etc.) the difference in the number of photons will be the radiometry associated with the aperture (the F# effects). So the aperture reaching radiance gets multiplied by the G# (essentially, the solid angle of the camera) to get to irradiance, which gets multiplied by the effective area of the detector array to get watts which are converted to photons. In summary, LIDAR model output is focal plane reaching flux (watts or photons) and normal is aperture reaching radiance.
The DIRSIG model actually computes fractional photons under-the-hood because it is really tracking bundles of energy that are weighted by reflectances and transmissions that are inherently fractional.
The output contains the steady state photon rates. Hence it is possible to get fractional counts. Consider this: if all counts are rounded down to zero for each of 100 bins and a detector was to integrate over all 100 bins, then the result would be zero photons.
In reality, the arrival rate for any single bin was less than one but every few bins it is expected to have at least one photon.
DIRSIG normally writes out the fractions and leaves it to the sensor model that processes the data to sort out the arrival statistics problem. It is possible to enable Poisson Statistics in the LIDAR capture method. In this scenario, each bin will reflect integer photon arrival counts. For the same simulation that generated the fractional counts, there will now be integer counts every few bins.
As clarification, the units are the same either way. Whether integer or fractional counts are used is purely conceptual. The fractional photon count just say clarifies the mean rate for a given time bin. Shooting the same pulse over and over should maintain the mean. With Poisson statistics, the integer value for the bin may vary (sometimes 1, sometimes 2, etc), but the mean should stay the same.