Documentation of using the code

[C-H-Chien]

Hi! I'd like to try on the code, but looks like there is no documentation for dependency installation, where the new user could start with, etc.
Would you by any chance kindly make a detailed documentation on your code? Any resources would be highly appreciated. Thank you!

[Astik-2002]

@C-H-Chien did you manage to run this code? For me, I keep getting the message "attempting to find map scale" in the ros mono node. The pangolin windows are blank, even though the images from both sensors are visible in rviz2

[C-H-Chien]

@C-H-Chien did you manage to run this code? For me, I keep getting the message "attempting to find map scale" in the ros mono node. The pangolin windows are blank, even though the images from both sensors are visible in rviz2

Unfortunately I was not able to run the code either. It was about half a year ago so I can't remember the details, but I guess I met roughly the same situation as you met... Eventually I gave up trying. The code however helps me better understand the methodology proposed in the paper.

[Astik-2002]

@C-H-Chien hey, so I got the code working (kinda). I think the problem is with the images we are getting from the simulator. I removed the sim loop and connected a realsense camera, then passed it's image to the ros_mono package. The orb slam part was working (pangolin window showed the camera image with orb features, and we had some tracking also). It's slow sometimes due to visualization. I'll remove pangolin and try to get vis results using just rviz

[C-H-Chien]

@C-H-Chien hey, so I got the code working (kinda). I think the problem is with the images we are getting from the simulator. I removed the sim loop and connected a realsense camera, then passed it's image to the ros_mono package. The orb slam part was working (pangolin window showed the camera image with orb features, and we had some tracking also). It's slow sometimes due to visualization. I'll remove pangolin and try to get vis results using just rviz

Wow fantastic! Terrific!!! Maybe you can open a PR or leave a detailed documentation for the future users to try on 😄 Have you tried not to connect a realsense camera but feed a sequence of images from the dataset?