PCA of Spatial Statistics Including Orientation
Long Term Goals
- Establish a data structure for the quantification of Poly-3-hexylthiophene (P3HT) thin film morphology
Determine Orientation
See An algorithm for segmentation of images containing non-overlapping fibrilar domains Bin the angles like so: Bin 1: 90º or -90º .. basically straight up or down Bin 2: -45º Bin 3: 0º horizontal Bin 4: 45º Bin 5: amorphous
Spatial Statistics
There are five micro states, a 5x5 matrix of spatial stats.
Began with independent stats, so the fist four elements in teh first coloumn.
- 1-1 autocorrelation
- 1-2 cross correlation
- 1-3 cross
- 1-4 cross
PCA
Autocorrelation of 90 degree fiber in the PC1
Cross correlation of 90 and 45 degrees in PC1
Cross correlation of 90 and 0 degrees in PC1
Autocorrelation of 90 degree fiber in the PC2
PCs vs Mobility
We see that the structure with the highest PC1 score has highest mobility.