Here you will find links to existing software (not mine) that might be of interest to people who work in vision, machine learning and robotics, especially for those who are doing systems integration in any of these areas. In no particular order:
fovis: fast odometry using vision Performs visual odometry based on a RGB-D camera, such as the Kinect.
gtsam: Georgia Tech Smoothing and Mapping Performs smoothing to solve the SLAM problem.
ethzasl_ptam: ETH Zurich Parallel Tracking and Mapping Improvement of the original PTAM for micro-aerial vehicles.
ethzasl_sensor_fusion: ETH Zurich Sensor Fusion An Extended Kalman Filter that fuses IMU and camera data to estimate the 6DOF pose of a micro-aerial vehicle.
rrt* Asymptotically optimal sampling algorithms.
bit* Asymptotically optimal sampling algorithm
deal.II Library for solving partial differential equations.
leda: Library of efficient data types and algorithms.
networkx: Library for handling graphs.
visilibity: Library for computing visibility graphs and polygons. I’m not sure if LEDA already supports this.
g2o: General graph optimization library. Very useful for bundle adjustment and SLAM types of computations, but much more general.
libcbdetect: Sub-pixel checkerboard detection for camera calibration.
libviso2: Library for visual odometry, both for monocular and stereo cameras.
psopt: Library for optimal control.
kalibr: Library for calibrating multi-camera and IMU sensor rigs. It’s great.
okvis: Monocular and stereo + IMU SLAM system
mcptam: Non overlapping field-of-view multi-camera PTAM
ros: Robot Operating System
theano: Python library for optimizing and evaluating mathematical expressions, such as the ones involving tensors in deep learning
tensorflow: Similar to theano
autograd: Python library for automatic differentiation
torch: Machine learning library for scientific computing and deep learning.