LSD-SLAM: Large-Scale Direct Monocular SLAM (ECCV '14)
Вставка
- Опубліковано 11 вер 2024
- Open-Source Code Available! see vision.in.tum.d...
Publication:
LSD-SLAM: Large-Scale Direct Monocular SLAM (J. Engel, T. Schöps, D. Cremers), ECCV 2014.
Authors:
Jakob Engel
Thomas Schöps
Daniel Cremers
Short Abstract:
LSD-SLAM is a novel, direct monocular SLAM technique: Instead of using keypoints, it directly operates on image intensities both for tracking and mapping. The camera is tracked using direct image alignment, while geometry is estimated in the form of semi-dense depth maps, obtained by filtering over many pixelwise stereo comparisons. We then build a Sim(3) pose-graph of keyframes, which allows to build scale-drift corrected, large-scale maps including loop-closures. LSD-SLAM runs in real-time on a CPU, and even on a modern smartphone.
![[ECCV 2024] Power Variable Projection for Initialization-Free Large-Scale Bundle Adjustment](/img/n.gif)
Looking forward to the presentation. Hopefully the live stream from eccv'14 will be a bit more stable than yesterday.
Looking forward to the release of the open source code!
It's released by now, in case you didn't see it.
@@DanHaiduc Do you know where the code is? I couldn't find the code base
do you have the resulting pointclout for this in ply format ?
I saw the other examples on your site but i need a large dataset as an example as i am trying to do something simmilar (with lsd slam) on a RPi V2 B
Great!
can I download android apk? I have lg g4 I think it could work
Great job guys !
So will this LSD SLAM be effective in a narrow hallway or in a straight metallic pipe where there are very less or no features?
Could you make a step by step video about setting the (LSD-SLAM on github) in visual studio? Thanks!
Very Impressive! I'm a newbie to SLAM. Did you use ROS in your implementation? I'm using ROS with Monocular ORB SLAM 2 and was wondering how you are doing your map visualization. Can you please tell me how you are displaying your LSD SLAM Map? Did you use RVIZ or RQT? Thanks for the help!
Did you ever find out how they did the visualisation?
No response yet from anyone yet! Are you working with any of the SLAM implementations ... LSD or ORB SLAM? I'm assuming that you have similar questions about how to visualize a map using monocular vision.
Yep, currently working on SLAM, was going to implement a few (ORB SLAM2, LSD-SLAM, OKVIS etc.) - I am going to take advantage of the IMU on mobiles too.
I want to use something like LSD-SLAM for drone mapping without processing the images afterwards for hours just to find out that I missed a spot..
But I have no idea how 😭 is there any program program that I can just download and use? My programming experience is limited to LEGO mindstorms, batch processing and Casio-basic 😑
I am wondering how was the visualization in 3d done? I mean what software package did you use?
In LSD-SLAM you are directly using video frames. Is there a possibility of data fusion from other types of sensors in addition to video to improve the trajectory estimation?
yeah, you can use a Kalman Filter for example and addd information from an IMU
I don’t get it..
I thought I am skilled with computer stuff, but.. I can’t get anything to compile. ANYTHING.
My coding knowledge ends somewhere between Casio-basic and Windows batch processing.
I just want a point cloud I can use for further processing without hour-long photogrammetry 😭
Are you using camera with global shutter?
Is that a model correction at 2:48?
+Eamonn McStravick I believe so. I think they're correcting the model by closing the loop
+Eamonn McStravick I believe so. I think they're correcting the model by closing the loop
wow!
So German English!!!!!!