Unnamed Aerial Vehicles (UAV) are known for their application in surveillance and tracking with on-board cameras. Videos from UAV usually suffer from jitter and high frequency unintended movements which makes necessary stabilize the footage. This is a Computer Vision problem, particularly difficult due to level of noise in acquisition and different types of scenarios which make hard to implement online stabilizer (on the fly). The present thesis explains in depth algorithms, methods and implementations of the OpenCV module videostab. We presented a version in C++ of videostab in context of video from UAV. Two main blocks constitute the current implementation: motion estimation and motion stabilization. In the former, the system robustly estimates global displacement in between every consecutive pair of frames. The latter can be used in two possible ways: Assuming that estimated motion contains intended UAV motion and high frequency vibrations or assuming that all estimated motion is undesirable. The first case is solved with a Gaussian filter to smooth motions and, in the second case, a filter called Zero Motion was implemented to give sensation of stillness. Results show that the implementation is robust and can work in different scenarios of UAV. Moreover, it is possible to run it in a general purpose computer with high speed performance. Online performance can be achieved using some function with graphic processing unit (GPU).