Saturday, November 7, 2009

Week 6 Day 3

Today I realised that block detection apparently failed for blue color. As the LAB's value for dark blue color is very close to the values for black, the application would sometimes regconize black as blue, which results in it detecting irrevelant objects in the surrounding as a block.

I have added codes to allow my program to ignore the environment's default LAB values, and it does provide a little improvement on the block detection for blue, but that was still not sufficient to allow blue block detection to work completely

Next, I expanded the range for block detection, and put a tighter limit for the initial detection point. This does helps in block detection for red, green and yellow, but blue is now not even detected.

For the rest of the days, I tried on expanding and changing the range for block detection of blue, which apparently still doesn`t work as it detects the surround's dark black as blue

No comments:

Post a Comment