As this project progresses... I came out with many difference names and this is the best one..
" Da Cam Whore "
" Da Cam Whore "
Sorry for not blogging, I am very busy with my schedules till I am very darn lazy to update my blog.
Now what I have to say here is that I had finally completed the first phase of this project and now what is left is documentation and some minor tweaks and optimization.
The hardest thing about this project is the calibration and rectification.
You have to make the cameras row aligned.
When it does, everything will go well.
I spent days on adjusting the BMSTATE parameters and it turned out to be calibration failure....
So... tons of calibrations have to be done...
This is the disparity map.
When an object is near to the camera, it is brighter than those which is further.
Strangely enough, what I got here is opposite... heck care...
as long as I have a correct output...
*bad engineering attitude ~.~*
Eventhough the disparity map is inverse, all I have to do is to change red to blue and vice versa..
simple right?
As you can see in the picture, the 3 cat dolls have different colour value which
indicates how far they.
Not too shabby for a beginner right?
The effective depth is 30cm to 250cm anything nearer or further is not that clear.. This is where phase 2 comes into action...
Above are some pictures of the rack which ma team has produced.
I stucked the cams with super glue after manually row align it.
This is very important in order to get a good depth map..
So... please take a good look at it, we are going to dismantle it and move on to phase 2
A little briefing on phase 2, it is called "multiple baseline awesome stereovision"
(for now)
Well, I will not update more on this topic due to some....
some.... ... ...
Dude... I need privacy and secrecy....
Everything from now onwards are my team's secret.
Ma manager has a strict policy on this =x
Before I end, I need to thank my supervisor mr.T and dr.T for giving us guidance and support...
Take care~