JetScan
JetScan is an Instant RGB-D 3D Scanner based on
- Jetson Nano Devkit by NVIDIA
- Intel Realsense D400 series depth camera
- Open3D by INTEL ISL
- theNded’s CUDA implementation
Summary :
This project was made in a motive to create metrically accurate RGB-D 3D scans , which is affordable by any creator / maker around the world.
Low power commodity RGBD depth sensors have been in the market for a while and they are also used in 3D scanning, also the aren’t completely portable. Custom portable 3D scanners cost a lot in thousands of $.
My system design is basically motivated by the problem of portable 3D scanning and making it cheaper and afforable and on the go processing on the 3D scanning module (edge computing) without requirement of external computing power, it gives the user a HD 3D model of the scanned object or environment instantly.
Workflows:
- RGBD sequence input → Making Submaps → Submap registration → Refine registration → Integration of final scene
Tested with Only CPU , Hybrid and Only GPU pipelines
GPU accelerated offline reconstruction in Jetson Nano itself was 15 - 20 times faster than only CPU pipeline.
Giving relatively good models with good voxel resolution in 2~4 mins (180 sec).
Detailed Code and Tutorial:
Future works include:
Implementing in Jetson Xavier AGX / Xavier NX
with multiple other RGB-D sensors like SR300 , Structure Core , Azure Kinect etc
Making it highly portable and real time online recontruction.
References:
http://www.open3d.org/
http://www.open3d.org/wordpress/wp-content/paper.pdf
https://github.com/intel-isl/Open3D
https://github.com/theNded/Open3D
http://dongwei.info/publications/open3d-gpu.pdf
https://github.com/IntelRealSense/librealsense
https://github.com/JetsonHacksNano/installLibrealsense