I am trying to find the best way to turn some representation of a line
based on 2 pairs of xy coordinates into ascii format in order to extract
habitat information associated with used versus randomly generated
animal movement paths.
I have approximately 33,000 observations (trajectories) to process and at
least 5 expanatory matrices to consult so I am necessarily seeking the most
parsimonious method possible. At present I have considered numerous
possibilities:
1. Using basic trigonometry, generate points along each trajectory at an
optimal uniform distance apart (roughly equal to pixel resolution) and use
"x" & "y" values to spatially locate the correct cell in the ascii grid with
which to populate a new object (my function, which seems to work fairly well
but takes time and difficult to handle if buffers are required).
2. Create lines in ArcGIS using start and end points, buffer lines, then use
"Thematic Raster Summary" in Hawth's Tools to generate a frequency table of
habitat types (also long, involves a lot of clicking and creates a lot of
files)
3. Using the xy points my function generated I can use buffer.line() in
adehabitat to iteratively turn a more-or-less continuous line of points into
a buffered pixelized line in ascii, but this still still a very
time-consuming process as each observation can take numerous minutes
depending on step length.
4. Using another method I haven't yet discovered, somehow iteratively create
a binary ascii grid for each observation using x1y1 and x2y2 which can then
be multiplied against the output grid based on the explanatory grid's
coordinates.
It should be stated that the reason I cannot just turn my entire dataset
into ascii binary format (0, 1) is that there is too much overlap in the
trajectories and I would only be able to retain one unique value per pixel,
thus losing valuable information.
One final option is if there were a way to import line shapefiles directly
from ArcGIS into R in ascii format, but this seems to only work for polygons
(readShapePoly) and likely points.
Any leads?
Tyler
[[alternative HTML version deleted]]
_______________________________________________
R-sig-Geo mailing list
R-sig-Geo@stat.math.ethz.ch
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
