loading
 
Lane & Obstacle Detection
Dan from Canada  [4 posts]
11 years
Hi everyone, I'm thinking of using this software for a project for a scale lane & obstacle detecting car.

The car would be able to navigate a track with several lanes and avoid obstacles and other cars on the track by switching lanes.

Would this software be suitable for my application? Would it be enough to track obstacles and lanes? What kind of system should I be looking at to run it on? What kind of camera would work well for this?

Any help would be appreciated.

Thanks!
EDV  [328 posts] 11 years
I think it can be useful in your project:
http://www.youtube.com/watch?v=Rbhd50XlnEk
http://www.sampson-jeff.com/roborealm/
Dan from Canada  [4 posts] 11 years
That's really helpful already, thanks. How are you interfacing between the Camera and the software? Or what is the software running on?

Thanks again
EDV  [328 posts] 11 years
Quote from [ http://www.sampson-jeff.com/roborealm/readme.txt ]:

There are examples for RoboRealm.

I used a Philips SPC900 CCD web camera. The CCD web cameras seem to work much than the CMOS imager cameras if you want to use them outdoors. A neutral denisty will help also.

One example shows sidewalk following. It uses test5a.robo and sidewalk1.wmv

The other example tracks an orange traffic cone at a distance of 50 feet. It uses files track_conea_3.robo and either of the uvs090427-007_50feet files. The .wmv is much smaller in size but is compressed and has drop outs in RoboRealm. Probably tweaking the values in RoboRealm will allow it to run correctly. The .avi file is directly from the camera if you want to see how it works with this camera. But the file is large (86MB, it is only 60MB zipped).

You can just view the results on YouTube here:

http://www.youtube.com/user/jws8675309
Dan from Canada  [4 posts] 11 years
So you are just running RoboRealm on a laptop then I guess and have the webcam connected to that as well?
EDV  [328 posts] 11 years
Sorry but it was not my experiment I just found it accidentally :)
But you could ask author of this work: http://www.youtube.com/user/jws8675309


However I can tell about AVM Navigator and I think that it also can be useful in this task ;)

The AVM Navigator makes easy a training process on new images:
http://www.youtube.com/watch?v=oRpXBOX5rxY

Also AVM Navigator provide several navigation templates such as:
*Follow me
http://www.youtube.com/watch?v=HTxNlOpm11U

*Walking by gates
http://www.youtube.com/watch?v=xbCpthKrL0o

*Navigation by map
http://www.youtube.com/watch?v=5qKQg5e4wbc
http://www.youtube.com/watch?v=Ggv8MaeJcTo

For more information see:
http://www.roborealm.com/help/AVM_Navigator.php
http://forums.trossenrobotics.com/showthread.php?4764-Using-of-AVM-plugin-in-RoboRealm
Anonymous 11 years
Dan,

Start by creating a video that you think best represents what your needs will be. Once you have that video (does not need to be very long) you can start to test various techniques. If you post it on the forum we can try a couple ideas to see which one would work best for your road conditions, camera type, etc.

Getting the video as close as possible to what you would expect the car to see is ideal. There are a lot of subtleties to vision which can create unexpected problems without realizing it.

STeven.

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