Does anyone have experience working with the Spectral Python module for land classification? The following k-means clustering example takes an iterative approach to classification through pixel clustering. Any thoughts, examples and/or visuals from your own work with SPy would be appreciated.
Source: SPy
>>> (m, c) = kmeans(img, 20, 50)
Using single-pass cluster algorithm to initialize clusters.
Clustering image...done
Starting iterations.
Iteration 1...done
Iteration 2...done
18743 pixels reassigned.
Iteration 3...done
3387 pixels reassigned.
Iteration 4...done
2512 pixels reassigned.
Iteration 5...done
1960 pixels reassigned.
---// snip //---
Iteration 28...done
149 pixels reassigned.
Iteration 29...done
135 pixels reassigned.
Iteration 30...done
107 pixels reassigned.
^CIteration 31... 15.9%KeyboardInterrupt: Returning clusters from previous iteration
>>>
Does anyone have experience working with the Spectral Python module for land classification?You might want to be more specific. In its current form, this question isn't really a good fit for the QnA format as it doesn't have a definitive answer. – R.K. Oct 29 '12 at 12:00