The purpose of this project was to use the Erdas Imagine Software to conduct a supervised image classification. The study area for this project was the
Annapolis Valley. The image used for classification was a Landsat 8 combined image taken on September 5, 2014. In addition students from the Centre of Geographic Sciences (COGS) conducted a “ground truthing” exercise in early October. Students visited a collection of field sites and took detailed notes describing the locations. Ground truthing locations are labelled on the Landsat true color image.
The purpose of this project was to use SPOT satellite imagery in an unsupervised classification. The SPOT satellite image was first orthorectified using secondary sources (national road and hydro vector files). The unsupervised classification was run on the orthorectified data, and then grouped into appropriate land cover categories. Validation files were provided to the students to assess their classification and to enable a quantitative analysis of the performance of various classification, grouping and filtering methods. After a satisfactory classification was complete, the image was transformed into a vector image for use within the ESRI ArcMap software package. Each student in the class selected a different ‘scene’ and the combined class covered almost all of Nova Scotia. An important part of this assignment was keeping data organized and cohesive so that all the classified images could be combined for a final project. The DEM and SPOT imagery was accessed through the GeoBase website http://geobase.com, an excellent source for free imagery.