Build a small desktop application

I need a Screen Change Detection application very similar to the Screen Change Detection feature found in the Auto Mouse Click software by MurGee.com. The feature in this software works by selecting an area of the screen to monitor, you click a start monitoring button and the application measures the percentage change in the pixels in the screen area being monitored. You can set a range of percentages so for example if the screen area changes by between 5 and 10% then an alert happens. Nothing happens when the screen area percentage changes by more or less than this range. I would like to make a small addition to this by adding a timer. The timer would start when i click Start Monitoring and then each time the screen area changes by the set percentage range the timer would record the current time. I would then want to export this column of time data to excel. Here's an example: - I select an area of the screen to monitor - I set the screen area percentage change parameters to between 5 and 10% - I click start monitoring - the timer starts - 4 seconds later the application detects a screen change of 3% - nothing happens - 1 second later the application detects a screen change of 7% - the current time on the timer of 5 seconds is recorded - 0.82 seconds later the application detects a screen change of 6% - the current time on the timer of 5.82 seconds is recorded - 1.11 seconds later the application detects a screen change of 8.5% - the current time on the timer of 6.93 seconds is recorded - 1.29 seconds later the application detects a screen change of 5.72% - the current time on the timer of 8.22 seconds is recorded - 0.01 seconds later the application detects a screen change of 12.3% - nothing happens - I click stop monitoring button and the data is exported to a column in excel like this: 5 5.82 6.93 8.22 The software needs to be accurate with its time measurements to within 0.03 of a second so it needs to have fast image processing capabilities. The sort of screen change i am trying to detect can be seen in the video i have uploaded (and the images).