Recording an objects position
mo from United States  [1 posts]
17 year
Hi im new to this program and haven't even cracked it open yet, but i had a quick question

i was looking at the green ball recognition tutorial and i was wondering if this was possible, to have a green ball and a red ball in a room, and have a random wandering robot. if it sees one of the balls in its wanderings it marks its location so it knows where it is to get back to it, and then when it sees the other it does the same thing and then finds out the distance they were apart

i cant think of way to mark the location, at first i thought gps some how, but i doubt its accurate enough or if its even possible, and since the robot is moving the distance cant really be relative to the robot itself.

how would you guys accomplish this

( for the record this is a brain storm, i don't plan on attempting this anytime soon as i have not even used the program yet)

anyways i haven't gotten to learn a new programming language in a while so i look forward to my new hobby :)
Anonymous 17 year
I am certainly no expert, but I like to pretend I know what I am talking about.

I think you would need wheel encoders (a way for the computer to count how many revelutions your wheels turn) assuming no slip one can then determine the distance you travel from some known starting point.  Your bot would then travel to one of the balls and you would acertain it's position relative to your starting point and the angle of the camera (if you aren't quite sure of the object's size)

I would also think you could modify an optical mouse to track movement along a smooth floor (if you are able to control that) and then you wouldn't have to worry about wheel slip.

Once you have the sets of coordinates for the balls it would be easy to get a distance between the two.
on second thought
17 year
Couldn't you remain in one place and make conclusions about the positions of the objects based on the servo positions as soon as it is centered in the screen and assuming the camera is mounted sufficiently far above the surface that the balls are on and then corraberate the data with their size (assuming it is known before hand)?
Anonymous 17 year
If you can put your camera on the roof, it is possible to track 2 or 3 objects location at 5 FPS. My plan is using a web camera with 1280X1024 resolution, careful aim the camera to the center of the field or corner of the field. then you can use COG module or bolb module to find out the relevent location. Good Luck.
Anonymous 17 year
for the wheel or optical mouse thing, how could i use that information to get back to the object or realize its distance?? i would know how far its gone but it most likely wasn't a strait line to the next ball so how do i turn that information into the distance they were apart??

thank you all for the replys

This forum thread has been closed due to inactivity (more than 4 months) or number of replies (more than 50 messages). Please start a New Post and enter a new forum thread with the appropriate title.

 New Post   Forum Index