Opencv on RTABMAP

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Opencv on RTABMAP

Nicolas
hello, first thank you for all you work, It has helped me a lot as several people more.
second, sorry for my english.

I make this topic because I have doubt with OpenCV version that can currently use . according to the installation guide is recommended to use the version 2.4.11 . But in another post you upload this link: https : //github.com/introlab/rtabmap/commit/6fe1e421816231f1f1a9345a1fe009975d43554b with support for version 3.0 OpenCV.

i'm very new in ROS. I have the version Indigo, with version 2.4.9 installed with indigo and another carpet name opencv. should remove these and then install version 3.0 or 2.4.11 from source as explained in this link??:
https://github.com/introlab/rtabmap_ros 
http://prog3.com/sbdm/blog/u013453604/article/details/49784351
http://prog3.com/sbdm/blog/u013453604/article/details/49781771.

I want to use find_object, and maybe make mi own contribution for the work im doin. i install from source, catkin for rtabmap_ros, and rtabmap without optional dependencies. funcionan perfecto.
sorry if it's a silly question, but I searched information, and also it is dumber be in doubt. thanks for your time and guidance.
Reply | Threaded
Open this post in threaded view
|

Re: Opencv on RTABMAP

matlabbe
Administrator
Hi,

This note about OpenCV in rtabmap_ros's build instructions is if you need the nonfree module of OpenCV:
If you want SURF/SIFT on Indigo/Jade (Hydro has already SIFT/SURF), you have to build OpenCV from source to have access to nonfree module. Install it in /usr/local (default) and the rtabmap library should link with it instead of the one installed in ROS. I recommend to use latest 2.4 version (2.4.11) and build it from source following these instructions. RTAB-Map can build with OpenCV3+xfeatures2d module, but rtabmap_ros package will have libraries conflict as cv-bridge is depending on OpenCV2. If you want OpenCV3, you should build ros vision-opencv package yourself (and all ros packages depending on it) so it can link on OpenCV3.
It is optional, as rtabmap and find_object can be built without the nonfree module. However, you will not have access to SIFT/SURF features.

If you want SIFT/SURF, you should rebuilt OpenCV from source. For that, versions 2.4.9 or 2.4.11 are fine. If you are going to use ROS, I don't recommend OpenCV 3 unless you are able to re-build all ROS packages depending on OpenCV 2.

cheers