I'm developing the use of UAVs out here in Kenya for wildlife and livestock census using vertical photography. Known as Aerial Point Sampling (APS), sample photos are taken at intervals along each flight line and the total number of animals is estimated from the numbers counted on the photographs.
The geographic coordinates of each photo are trivial to capture from the GPS systems. What is more difficult is the height above ground level (height AGL) which is required to exactly scale each photo.
My UAV (based on the MAJA airframe) maintains a pre-programmed QFE height AGL, and while to some extent you can pre-programme a height AGL flight profile it is still difficult to calculate the exact height AGL for each photo.
I have now taken delivery of a laser altimeter (71 grams) but at the height AGL I really want to fly (≈100m) its records are a bit flaky. Nonetheless, I am getting heights AGL for each X-Y photo coordinate.
I'm now trying a different approach which is to compare in real time the height AMSL from shuttle missions with the height AMSL for the GPS (the Z function). This is easier than it sounds, primarily as I'm not yet quite sure how quickly the GPS Z function can react to a small (say 5m) change in altitude.
But what I really want to do is to be able to programme into the ARDU pilot a flight profile so the UAV will maintain, say, an average height AGL of 100m +/- 10%.
Anyone out there working on similar problems? I know that cruise missiles probably do this in their sleep but we're not quite in that league!