Benjamin

I added a speeded up function `loadFromTextFile2` for reading and parsing an updated ORBVocabulary to my fork of ORB_SLAM2: https://github.com/Beewe/ORB_SLAM2_Windows/commit/bcd957c48c9ebc3bd4839bb5e9e036783faabd4f

I'm running my build on Win7 (but with VS2013) and can easily process more than 30fps on my PC and around 25 on my Laptop (also VS2013), so it may...

Yes, I've already have it running under windows. I'm planning to release this version on github, but first I have to merge the recent updates from this repo to mine...

Hey, as you've maybe already seen there is a new version called ORB_SLAM2, which includes several new features, so I decided to make this Windows compatible. Check out my fork,...

Just having a guess: Maybe you use an outdated version of Eigen, see "Prerequisites" from the install instructions; you have to use Eigen at least in version 3.1.0. I just...

You always need the Vocabulary file, it is used for place recognition and feature matching. There is already one in the Vocabulary directory which you you first have to extract...

Hello, I made some quick tests with both ORB_SLAM and ORB_SLAM2 with the videos "out640.mp4" and "output_480.mp4". With increased `ORBextractor.nFeatures` to 2000 I was able to initialize with ORB_SLAM, whereas...

No this resolution should be fine, in all my tests I use exactly this resolution, but I have no experience in stereo camera setup / stereo SLAM. Note: With increasing...

Much easier ;) the Tracking-Method already returns the camera pose. so the only thing you need to add to your example is: `cv::Mat cameraPose = SLAM.TrackMonocular(im,tframe);`

The camera pose you get is indeed not in world coordinates, you will need the inverse. So you basically get [R T;0 1] and the inverse would be [R^t -R^t*T;0...