How would the camera determine depth and object avoidance?

The camera would scan and know where the objects were. Then it would look for a clear path in front or to the right or to the left, etc. With the Kinect it would be much easier, but with a regular webcam people are doing it in Roborealm and other programs.

With motion you can determine some distances I guess. But it's not very environmental friendly. For a vision processing only robot, roborealm is great. It might be a method I can adopt, but not right away. Combining a distance sensor with the camera is my personal preferred method. But who know, I might wake up tomorrow with a brilliant idea or not!

You have been known to do that. If I put sonars, I will have to have quite a few. Because the robot will run into a table that has an empty space below. Or a shelve above. so, I have to cover top and bottom and left and right. Of course a servo will cut down the number of sonars. But, sonars are blind to some things, so I will also have to have IRs. I have plenty of IRS now. But, they don't see black things. It would be so simple if we could just use the picture from a camera. One person did this. He put TWO cameras with automatic focus. If camera A was in focus, then the item was at a distance. But, if camera B was in focus, it was right close by. I don't know how he programmed it. But, that is some food for thought.

I am not a very good programmer, so I will rely on help from others. I am very proud of your progress with the EZ-Board and software. I look forward to the April Upgrade.

Thanks,

Mel

DJ Quote "I might wake up tomorrow with a brilliant idea or not!" LOL! JW :)

@moviemaker in the meantime you can use two hc-sr04 distance sensors. one horizontal, and one vertical. That's how my K-9 works.

My robot is about 4.5-5' tall. do you think that Only two scanning sonars will do the trick? I don't have to Really know the EXACT distance, only that SOMEthing is in the way, DON'T Roll over it. Just to stop and avoid Obsticals.