I'm currently using goodFeaturesToTrack and I'm getting good results. The points are in a goofy order so I'm deciding how to order the points or if I'm better off making my own point detection method. I'm thinking it's best to use the points and sort them rather than finding them myself. I think it would take me half the day to make an algorithm like that and I'm running out of time.
For clarification cvFindDominantPoints requires cvarr etc, not the normal Mat c++ type format. I'm trying to find an actual supported method for finding the defining points of a series of connected lines.
Imagine making straight scribbles on paper. I want to find all the corners. Granted, it's pixels so a straight 45 degree line could be made of 100 corners. dominantpoints addresses this issue but I can't find how to do it without dominantpoints.
FYI I now have the points in order. At this point I'm trying to get a warp perspective done so I don't have to try to figure out the angles with my own math.
Wrapping up calculating the path and getting ready for dead reckoning etc. I solved my segmentation faults using gdb. Turns out an algorithm I'd copied had a little goofiness to it.
I am planning on making a tutorial based on this soon. It currently works but I want to add features and optimize the code a bit. Is anyone interested in working in parallel with this project? I'm interested in making a github collaborative project to make an impressive Prop/Pi Bot.
It would be great to turn this in to a collaborative effort. I want to get the bot balancing but Andy thinks it will be tough with the ActivityBot straight from the box (which I then add wood to to shift the CG). I almost got it balancing but had to put that on the back-burners to be ready in time for Maker Faire Detroit.
The robot worked great, when it worked. Unfortunately about 100 people tripped over the bright red huge blanket and calibration is a bit sensitive to the blanket being the right size and shape. I'm glad I could use this to get kids interested in STEM and help promote Parallax products.
Let me know if any of you are interested in tweaking this on GitHub.
Comments
Imagine making straight scribbles on paper. I want to find all the corners. Granted, it's pixels so a straight 45 degree line could be made of 100 corners. dominantpoints addresses this issue but I can't find how to do it without dominantpoints.
cvFindDominantPoints is deprecated... Several sites mentioned that using approxPolyDP is the closest built-in method to cvFindDominantPoints. This example is not exactly what you are looking for, but it explains the use of approxPolyDP: http://docs.opencv.org/doc/tutorials/imgproc/shapedescriptors/bounding_rects_circles/bounding_rects_circles.html
I hope this helps.
dgately
FYI I now have the points in order. At this point I'm trying to get a warp perspective done so I don't have to try to figure out the angles with my own math.
Video of the ActivityBot solving the maze soon.
https://www.youtube.com/watch?v=stuCPNpmuDA
It would be great to turn this in to a collaborative effort. I want to get the bot balancing but Andy thinks it will be tough with the ActivityBot straight from the box (which I then add wood to to shift the CG). I almost got it balancing but had to put that on the back-burners to be ready in time for Maker Faire Detroit.
The robot worked great, when it worked. Unfortunately about 100 people tripped over the bright red huge blanket and calibration is a bit sensitive to the blanket being the right size and shape. I'm glad I could use this to get kids interested in STEM and help promote Parallax products.
Let me know if any of you are interested in tweaking this on GitHub.