h9v7

The area of my focus is in Central America from Southern Mexico down to Nicaragua. The Global Change model showed large area of negative changes in the southern Mexico region and a large area of positive change Honduras and Nicaragua. My goal is to see why there is such a large change from one side of the area to the next if it is because one area is better for farming or our there political/ecomomic policies that are causing this outcome.

=__**﻿EVI's**__= I used an Enhanced Vegetation Index(EVI) to be able to distinguish between the different types of vegetation in the area. I choose EVI over Normalized Difference Vegetation Index(NDVI) because of the worry about saturation. From the different time frames using the same classification ranges you can see the changes in areas that are believed to be major farming areas for the region. On the northern coast of southern Mexico is the best example because it ranges from dense vegetation to relatively barren land depending on the time of year. It is also the area that showed the most negative changes on the original change detection map.
 * January 17 2010 May 09 2010 September 30 2010**

=__﻿List of Available Images__=


 * 2/28/2011 || 5/2/1999 || 3/25/1997 || 4/5/1995 || 4/15/1993 ||
 * 4/21/2001 || 2/27/1999 || 2/21/1997 || 3/20/1995 || 3/30/1993 ||
 * 4/5/2001 || 2/11/1999 || 5/9/1996 || 2/16/1995 || 2/26/1993 ||
 * 10/27/2000 || 4/13/1998 || 3/22/1996 || 1/15/1995 || 2/10/1993 ||
 * 3/1/2000 || 3/28/1998 || 3/6/1996 || 12/30/1994 || 1/9/1993 ||
 * 2/14/2000 || 2/24/1998 || 1/18/1996 || 11/28/1994 || 3/30/1990 ||
 * 1/13/2000 || 2/8/1998 || 1/2/1996 || 3/1/1994 || 4/25/1988 ||
 * 12/12/1999 || 1/7/1998 || 5/23/1995 || 1/28/1994 || 3/11/1986 ||
 * 9/7/1999 || 6/13/1997 || 5/7/1995 || 11/25/1993 ||  ||
 * 5/18/1999 || 5/12/1997 || 4/21/1995 || 5/1/1993 ||  ||

=﻿__Processed Images__=



=**__Filtering__**=

These next image are all the same image using two different filters and two different kernel sizes for each filter. They all show the same city in the image to better show what each filter can do. I also used 60% image add back as a standard amount for all the filters.

(From left to right) High Pass 3x3 **ll** High Pass 7x7 **ll** Median 3x3 **ll** Median 7x7

=**__Principle Component Analysis__**=

Next I used Eigenvalues to determine which bands had the information I needed and which were simply just distortion. As you can see from the graphs and tables below you can see just how much information is stored in just the first band alone. The second and third band included additional information but as you will see in a later image the third band still included some distortion so for the final images I only used the first two bands for the analysis. 2000 Eigenvalues


 * 1 || 3009235.377 ||  || 0.922013 ||
 * 2 || 180627.4597 ||  || 0.055343 ||
 * 3 || 52963.30218 ||  || 0.016228 ||
 * 4 || 11696.15123 ||  || 0.003584 ||
 * 5 || 6259.751455 ||  || 0.001918 ||
 * 6 || 2986.013124 ||  || 0.000915 ||

2011 Eigenvalues

The image has a lot of the original distortion removed but some still remains. The streaks in the water are a good indicator of the distortion. The next two images are the final images after using only the first two bands which contained 96-98% of all the information in each of the images.
 * 1 || 2007327.209 ||  || 0.91645 ||
 * 2 || 107234.7068 ||  || 0.048958 ||
 * 3 || 48355.20045 ||  || 0.022077 ||
 * 4 || 12433.01353 ||  || 0.005676 ||
 * 5 || 9487.266458 ||  || 0.004331 ||
 * 6 || 5492.600838 ||  || 0.002508 ||

These images clearly show the area of focus for me which is the swampy region on the right side of the images and is the largest concentration of negative change from previous studies.

=**__Change Detection__**=



=**__Supervised Classification__**=






 * Region of Interest Separability Tables**


 * Confusion Matrix for Images**


 * Report**