Personal learning excursion into Simultaneous Location and Mapping (SLAM) methods. What's currently there is in the tradition of the ORB-SLAM family of systems. However, as a learning exercise, I decided to implement the algorithms from rather low above ground up. So, for example, there's my own implementation of ORB feature detection, description and matching on top of the Boost Generic Image Library (GIL) as image data structure, or a Pose and Places (PnP) optimizer using least squares just on top of linear algebra operations using the Eigen library.

I am focusing my efforts on RGB-D camera input, such as coming from an Intel RealSense camera. This is because I own such devices.

I have been using CMake and Vcpkg to manage dependencies and build the the system. In addition, you will want to get a copy of the LORIS data set (see Data Files below).

Warning: I am still in the middle of completing/reworking the initial version, so maybe hold off for a while until you try this code. Or shoot me an email.

In order to execute tests and example, a data folder needs to be created. Specifically, the folder should contain the OpenLORIS Scene Data Sets.