READ ME
Gavin McNicol - EPS 209 Matlab for Earth Scientists

This file explains the Final Project submission.

The code requires the Image Processing toolbox.

The files submitted reflect the state of my work on this project at
the point that I presented on the 27.04.2011 although I intend to 
develop these ideas more fully. 


There are two main sub-directories in my submission:
(1) code_imageset1
(2) code_imageset2

In each of these are three items:
(a) An ROI.png file that delineates the relevant area of the image.
(b) A set of .jpg images.
(c) The MATLAB code file: FindCows_gavin
(d) The MATLAB code file: FindCows_gavin_color

In code_imageset1 (1), the images are a subsample used to infer
a threshold value for mean intensity that can be used to determine
if cows are present or not.

In code_imageset2 (2), the images are a full set of images from Day1-Day60
of the year 2010. 

(c)
There are two parameters that can be adjusted in the IF statements toward
the end of the code. These two parameters are the threshold values of
m and y, where m = mean intensity of true color image, and y = mean intensity
of range filtered image.

I have defined m > 100 and y > 3.5 based upon observations from running sub
sets of the images.

(d)
This is the code for the Mahalanobis distance which I perfomed and wrote but 
did not create thresholds for. I may incorporate this for detection.

NOTE: An vector of x values for the figures is created by a for loop in both (c) 
and (d). There were 24 images per day, so the 'Day of Year' is calculated by 
dividing the image number by 24, such that the x-axis on the figures
is more easily interpreted.