Separation of Carbonates by Using PCA on ASTER Bands

Figure (c):-PC4 in PCA analysis That indicating dolomite Dolomite as bright pixels.
For limestone separation from the other formations PCA analysis was performed on 4,6,8,9 SWIR ASTER. The following table (2) shows the eigenvector statistics of these bands. As concluded from this table, limestone is shown in -PC3 in bright pixels (figure d).
Table (2): Eigenvector statistics for ASTER bands 4, 6, 8 and 9
| | PC1 | PC2 | PC3 | PC4 |
| Band 4 | 0.452 | 0.791 | 0.058 | 0.407 |
| Band 6 | 0.583 | 0.052 | -0.428 | -0.688 |
| Band 8 | 0.500 | -0.292 | 0.809 | -0.101 |
| Band 9 | 0.453 | -0.534 | -0.399 | 0.592 |

Figure (d): -PC3 in PCA analysis That indicating limestone as bright pixels
Accuracy assessment of the research
The geological map which was complied from this area, confirms the high accuracy of this analysis figure (a). Dolomite appears in red figure (a) is in accordance with figure (c). Also pure limestone in dark blue and mixed limestone in light blue have been shown in figure (a), that are in high accordance with figure (d).