Improve channel matching algorithm

[lahiiru]

Step 1
Run the script image.py. It used an image from double_exposed_webcam directory. It first extract red and blue channels then calculates distance difference using template matching.

Step 2
Try changing parameters like white_threshold, grid sizes, template ratio, template matching algorithm etc. Ensure the displacement given by the grids using AmpedFive. E.g. 7.jpg has about 31 px difference among each channels.

Step 3
Think and develop an algorithm to match those channels robustly. You may apply filters before template matching. Further you can try any method other than matching.