Giter VIP home page Giter VIP logo

Comments (2)

aaguiar96 avatar aaguiar96 commented on August 27, 2024

Hello!
About the README you're write. We will fix that issue.

Question 1: For you to use the perspective camera in the omnidirectional version you'll have to calibrate it using the omnidirectional camera calibration toolbox mentioned in the README file. The SVD error is due to the fact that your are not using any calibration file. You have to calibrate the perspective camera and provide the file path into the launch file.
However, there is no advantage on using the perspective camera in the omnidirectional version. This version intends to solve the high radial distortion present in dioptric cameras...
You should use this version with dioptric cameras. To test it you have a dioptric camera sequence and a calibration file suitable for this camera.
You just have to set the correct path to the file into the launch file.

Question 2: You can try to tune the parameters in order to get a better performance. However, the conditions where the bag file was recorded are challenging (slow motion and high turbulence) and as this is a monocular VO method, the scale calculation is an inherent issue.

Hope this helps!

from viso2.

ysong-aceinna avatar ysong-aceinna commented on August 27, 2024

Hi @aaguiar96 , thanks for your responding!

For question 1: yes you are right, it works normally after I set the correct calibration file path in launch file, and I got trajectory from rostopic "/mono_odometer/odometry" as below. Similar big drift as I mentioned in question2.
Screenshot from 2019-06-24 08-52-04

Except monocular VO method, I also tried the stereo VO method and met similar big position and attitude drift which mentioned in issue 68. Except the test.bag developer provided, I also made a few bagfiles which generated from Kitti datasets to test the stereo VO method, when I found the big drift issue, I referred to a close issue 45 and set
~ref_frame_change_method = 2 ~refinement = 2
, but I didn't get an obvious performance improvement.

I see the scale calculation is an inherent issue of algorithm but not the wrapper, it is also greatly appreciated if you have any parameters suggestion to improve the stereo VO performance according to your experience.

Thank you very much for your input.

from viso2.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.