Procedure

By using Agbogbloshie as our test site, we approached this study by attempting to identify the spectral signatures of various E-waste contaminant attributes, and experimenting with various methods of remote sensing analysis. We looked at imagery from the visible and thermal bands, as well as created NDVI's to examine reflectance of live green vegetation. We also supplemented this with GIS vector data and ground level photo imagery, so as to spatially analyze and organize our findings as well as understand the damage of e-waste on this site at a closer level.
 
First, we did aerial photo analysis of Agbogbloshie using Google Earth Geoeye and Landsat imagery, to see if there were any identifiable contaminants or geomorphological effects of e-waste at this level. We also created unsupervised classifications of the dump-site, as a means of observing any uniform patterns in different contaminants. After observing a significant decrease in vegetation from the visible band, we created NDVI's for a series of Landsat images between the years 1985 and 2011 to create a measurement for the loss of vegetation in this area over time. Using the same Landsat imagery, we then created a change-detection in the thermal, to see if e-waste and its effects caused a general increase in temperature in this area. Finally, to see if the polyurethane fires common at this site can be identified using remote sensing, we examined Google Earth Geoeye imagery of the site.