 |
| Supervised classification of Angkor Wat, Cambodia based on SWIR composite. |
 |
| Google Earth image of potential pyramid locations in El Mirador. Potential sites, difficult as they are to see, are in red. |
 |
| Google Earth image of various layers of El Mirador overlaying the standard Earth view. The NDVI layer is visible. |
 |
| Google Earth image of El Mirador supervised classification overlay with legend. |
The final module of our initial project involved two main tasks. First, we were to export the layers created in the first two modules into a format that could be displayed and shared in Google Earth. This was a relatively simple task using the 'Layer to KML (Conversion)' tool for individual layers and the 'Map to KML (Conversion)' tool for entire data frames. The results can be seen above. I was disappointed in the appearance of my potential pyramid layer and some of the colors of my classification layer are off, but in general the process was a success.
A second task required of graduate students was to repeat the supervised classification process with Angkor Wat, a well-known archaeological complex in Cambodia. We were to locate and download appropriate Landsat imagery, then classify the image based on a false color, NDVI, or SWIR image. As La Danta pyramid was the focus of our El Mirador classification, the monumental core of Angkor Wat would be our focus here. The difficulty resided in the different land cover types that are included within the core. Every training sample I tried over-represented the likely locations of additional remains. However, my best classification restricted the core sample to the vegetation overlying the structures.
This was an enjoyable module that would have been more so if my computer had not expired in the middle of the assignment. Even so, I believe I was mostly successful in completing the two tasks. As always, however, more time could be spent on tweaking the training sample for a more refined classification.
No comments:
Post a Comment