Page Author: [Shubodh Sai](mailto:p.saishubodh@gmail.com?Subject=The%20Notion%20Note-taking%20Project%20|%20SLAM%20Frontend)
Source A
Everything in the context of SLAM.
The VO estimates a rough camera motion based on the information of the adjacent image, providing good initial value to the back end.
The VO algorithm is mainly divided into two major categories: feature point method and direct method.
We will only be considering feature point method for the advantages of stability, efficiency, insensitivity to changes in environment.