search

luigifreda / pyslam

pySLAM is a Visual SLAM pipeline in Python for monocular, stereo and RGBD cameras. It supports many modern local and global features based on Deep Learning.
GNU General Public License v3.0
1.91k stars 338 forks source link

Which code file to start with just for computing VO from monocular images? #68

Closed Mechazo11 closed 2 years ago

Mechazo11 commented 2 years ago

Dear @luigifreda,

Thank you for publishing your work as open-source. I am relatively new to the field of VO and currently working on a novel VO system that incorporates semantic information for place recognition.

A part of my system can compute key points and their descriptions using ORB in consecutive images using OpenCV.

My question are:

  1. Which scripts from your code repository should I start with that would compute VO given two sets of key points from two images (say t_0 and t_1)

  2. You mentioned ORB was improved upon in the ORB2-SLAM paper. Are those improvements already part of OpenCV or are they user-defined functions found in the python scripts in your repository? I am using OpenCV 4.2

  3. Is the camera calibration code available in your repo?

Thank you for your time.

With best, Azmyin

luigifreda commented 2 years ago

Hi, first, please take a look a the README. You can get there a reply to 1 and 3.

1 Screenshot from 2022-03-18 18-27-37

3 https://github.com/luigifreda/pyslam#camera-settings

As for point 2, no, that version of ORB (from ORBSLAM2) is NOT part of OpenCV. Hope this helps.