Opened 14 years ago

Closed 10 years ago

#1281 closed defect (fixed)

i.cluster not working with a high sample count

Reported by: nedhorning Owned by: grass-dev@…
Priority: normal Milestone: 7.0.0
Component: Imagery Version: svn-trunk
Keywords: i.cluster, cluster sample Cc:
CPU: x86-64 Platform: Linux

Description

I am using i.cluster on GRASS 7.0 on a 64-bit Ubuntu computer.

If I use the default values for "sample" and "min_size" i.cluster works but if I set "sample=5,5" the resulting signature file only has one class. When I enter "sample=5,5" I assume that every fifth pixel from every fifth line is being sampled for use in calculating the cluster statistics. This is apparently too many samples for i.cluster to handle. If I use "sample=100,100" the program appears to work. I have a feeling that the sample array is not able to handle a large number of samples but have not verified this. It would be nice to be able to use "sample=1,1" even on large images.

Change History (2)

comment:1 by martinl, 14 years ago

Component: DefaultImagery
Keywords: i.cluster added
Version: unspecifiedsvn-trunk

comment:2 by neteler, 10 years ago

Resolution: fixed
Status: newclosed

I have made a test with the example in i.cluster (using r61736) and did not run into this problem. Likely some internal data structures have been updated since your report.

I take liberty to close the ticket, feel free to reopen with an example (ideally based on the NC sample dataset).

Note: See TracTickets for help on using tickets.