| Author |
Message |
|
| amith4455 |
 Posted: Thu Feb 11, 2010 6:28 pm |
 |
|
|
Roboteer
Joined: 09 Feb 2010
Posts: 1
|
We are trying to do our final year project in Vision TRIx.
we have a concept in our mind and need ur help to do this
The Project we are thinking about is,,
There will be a start and end points, in between there will be obstacles,the bot should not hit the obstacles and go to the finish point and i should come back to the start point(in the same path which it traveled to finish point).
 |
|
|
| Back to top |
 
|
| rahul2047 |
| Posted: Sat Apr 24, 2010 3:42 pm |
|
|
Roboteer
Joined: 24 Apr 2010
Posts: 4
|
Hi there,
Your problem statemente can be solved by several ways. But if you want to solve it by robotic vision:
It is must that to startup with robotic vision, color based object detection should be used. Keep different colors for objects having different characteristics.
Suppose this is the image captured by an overhead camera mounted over the arena.(Assume this as input image)
you can see that according to the problem statement I kept the color red for all the obstracles, which are the areas where bot should not move and so on.
Correspondingly some images are shown here which will be the output as binay image for your correct thresholding of the image captured by overhead camera.
For Bots position and orientation:
You need to keep the bot's top plate with two diference colors. The centroid of those binary images will give you two points. By which a straight line can be drawn. then the slop of the the straight line can be known .. and this is your bot's orientation in the map. By which your program can deside which is forward for your bot and so on.
Start Point/Area
End Point/Area
Area where bot can move
Obstracle Area
Bot at any instance
At any instance the way of getting its position, its orientation and its distance from present goal keeping obstracles in account.
This is fully iterative way of doing ur work done.
Note: At any instance if bot knows only three points:
0. Map
1. Bot's position in that accounted map.
2. goal's position in the same map
.. any kind of problem statement can be solved.
Further for getting back at the start point again:
Keep record of all positions(co-ordinates in image) where bot has moved.
This will get back as the same way as it has gone to the end point. To make it smooth you can look for smoothing the path (line/curves) walked by the bot.
Possible binary image output in this kind of algorithm:
If bot walks this path:
The next movable area can be extracted by erod or any other suitable process and output can be like this:
Give it a try.
When you are done with this you can enhance your algorithm for precalculated path or better than this one. |
|
|
| Back to top |
|
|
|
All times are GMT + 5.5 Hours
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
