| Re: satellite field of view pole projection [message #93281] |
Thu, 02 June 2016 11:42  |
penteado
Messages: 866
Registered: February 2018
|
Senior Member
Administrator |
|
|
Hello,
If I understand correctly your problem, since you have the corner points, a solution might be to use pp_drawsphericalpoly from my library (http://www.ppenteado.net/idl/pp_lib/doc/).
Se the 5th example plot in http://www.ppenteado.net/idl/pp_lib/doc/pp_drawsphericalpoly .html, it is the one with polygons spanning the pole.
Paulo
On Thursday, June 2, 2016 at 7:22:27 AM UTC-7, audrey.sch...@gmail.com wrote:
> Hi everyone!
>
> I thought I had a rather simple (and common) problem, but I just can't seem to find the hang of how to solve this:
>
> I have a satellite that is flying around the Moon in a circular orbit with fixed altitude. The instrument on board the satellite is nadir pointing, has a fixed field-of-view, the surface projection of which is 150 x 25 km^2 (you can imagine the instrument being fixed in space and the Moon turning beneath the instrument). I now have a dataset of measured counts and associated longitude x latitude pair (either as boresight point, or as an array of four corner points).
>
> What I would like to do is map my counts onto the footprint of my rectangular field-of-view (around the boresight, or spanning the four corner points).
>
> This works well for the equatorial region: I just create a 360 x 180 float array, and for each measurement distribute the counts from the min(corners_longitudes), min(corners_latitudes) to the max(corners_longitudes), max(corners_latitudes).
>
> At the poles this obviously does not work, because at the poles a rectangular field-of-view projection has non-consecutive longitude, latitude values (they kind of jump around in value if part of the field-of-view is beyond the pole). I assume I have to do some triangulation, and fill all triangles within the triangles associated with the corner points. But I just can't seem to figure out how to do this.
>
> Does anyone have a solution to this problem?
>
> best,
> Audrey
|
|
|
|
| Re: satellite field of view pole projection [message #93282 is a reply to message #93281] |
Thu, 02 June 2016 23:27  |
audrey.schaufelberger
Messages: 10
Registered: June 2014
|
Junior Member |
|
|
thank you, that looks exactly like what I want to plot, but unfortunately I seem to already fail a step earlier in the process if I use your approach.
In your example you have a certain number of field-of-views you want to plot, which are not overlapping. In my question I was only talking about one orbit, and in that case your solution would work well.
But I actually have hundreds of orbits, where the fields-of-view overlap. I would like to average these measurements before I plot them. I am not sure how this actually influences the mapping process...
So far (for the equatorial region) I have simply been adding up the counts in one array, and have been creating a second array where I sum up the 'exposure' (how many times a surface element was seen). I then divided the first array by the second, the result of which I plotted.
If I wanted to use your approach, I would like to make use of the part where I determine what region a rectangle covers, but would then need to somehow be able to further process the result, and not directly plot it.
Any ideas?
best,
Audrey
|
|
|
|
|
|
| Re: satellite field of view pole projection [message #93361 is a reply to message #93282] |
Thu, 23 June 2016 18:21 |
penteado
Messages: 866
Registered: February 2018
|
Senior Member
Administrator |
|
|
The functionality to handle overlapping polygons, keeping track of everything that falls on each point in the map, so that one can decide on how to combine these values (take a mean, remove outliers, choose the stacking order, etc.) was lurking in pp_drawsphericalpoly, but because I had not decided on the algorithm and API, I had not documented it. I made some changes when I used it last week, and now I added some documentation, including some examples, of how to handle that (most relevant keywords are do_stack, stackmap and stackcount). I will probably add some alternative algorithms in the future, which would be chosen using other values for the keyword do_stack (at this time the only usable one is do_stack=1).
On Thursday, June 2, 2016 at 11:27:20 PM UTC-7, audrey.sch...@gmail.com wrote:
> thank you, that looks exactly like what I want to plot, but unfortunately I seem to already fail a step earlier in the process if I use your approach.
>
> In your example you have a certain number of field-of-views you want to plot, which are not overlapping. In my question I was only talking about one orbit, and in that case your solution would work well.
>
> But I actually have hundreds of orbits, where the fields-of-view overlap. I would like to average these measurements before I plot them. I am not sure how this actually influences the mapping process...
>
> So far (for the equatorial region) I have simply been adding up the counts in one array, and have been creating a second array where I sum up the 'exposure' (how many times a surface element was seen). I then divided the first array by the second, the result of which I plotted.
>
> If I wanted to use your approach, I would like to make use of the part where I determine what region a rectangle covers, but would then need to somehow be able to further process the result, and not directly plot it.
>
> Any ideas?
>
> best,
> Audrey
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
