scrapbook material for ya!
http://hackedgadgets.com/2009/02/14/vision-object-detection-and-ranging/
Thanks Rik!
Thanks Rik!
lol
Took you four months to come up with that snap eh?
Very well donethis is an
Very well done
this is an amazing project but i just have some questions. they may be very silly questions but they just popped into my head and i’m too lazy to analyse them or i don’t fully understand your system.
1). you said that you used cheap cameras but if you got better ones, then the software should work better. bu if you got better cameras, wouldn’t it use way more processing power because of the increased frame rate since there would be an increase in the number of analysed frames?
2). and also, better cameras would yield much more detail, and the problem i saw in your video of reflections would be much worse with better cameras. unless you add some filters i suppose. or am i wrong?
3). what would happen if you had the “machine view” of a light bulb (switched on), an empty glass, and an incline ( a slope viewed from the bottom going up).
keep up the good work though. very good.
I specialize in Silly
1. Whenever I refer to “cheap” cameras I’m referring that the quality of sensors versus other sensors makes webcams extremely “cheap” for the amount and variety of data they have to offer. Also there is a way to scale the image called “Pyramid Down” which allows lower resolution when you need it. If given a choice I would always go with the higher resolution quicker framerate webcam - If the software is good, it can throw away data if it gets overwhelmed.
2. All camera are different. The software has to be more intelligent in the future to unify objects. This would include smart filters.
3. it would be blinded - unless the camera had an auto-iris - or there was smart auto light balancing filter, all of them would be problematic -
At this times webcams are pretty good eyes, but we need a good visual cortex to make the data from the webcams useful