Skip to content

• Leveraged sequential LiDAR point cloud across 1800km, spherical view images & geolocation for analysis. • Implemented MRCNN, YOLO & performed Transfer Learning using MSCOCO for segmentation with 94% IOU. • Obtained 97.8% accuracy on point cloud clustering using DBSCAN, MeanShift, K-Means and PointNet

Notifications You must be signed in to change notification settings

Vidya1899/UFAnalysisCV

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Leveraging LiDAR and Street View Data for Road Feature Detection with Ordnance Survey Northern Ireland

Description

The National Mapping Agency of Northern Ireland’s (OSNI) mission is to provide high quality geospatial data. Historically this has been 2D mapping, but modern survey techniques and increasing user requirements have shifted focus toward 3D data. Since 2019, OSNI has operated a vehicle mounted Mobile Mapping System (Leica Pegasus) across Northern Ireland capturing 3D point cloud data and spherical street view imagery. This project seeks to explore the potential of this highly detailed LiDAR and Imagery data via machine learning and other data science methods, with a focus on developing pipelines to classify and identify urban features like drainage. We also welcome participants who are up for exploring various ways to process, study and visualise LiDAR and Imagery data of this detail.

Useful skills: Computer Vision (CNN), Geospatial Visualisation, LiDAR Data Processing, Machine-Learning

Contents:

  • Notebooks - contains jupyter notebooks to get you started
  • Whitebox tools contains code to install and run Whitebox in the DSG virtual machines

About

• Leveraged sequential LiDAR point cloud across 1800km, spherical view images & geolocation for analysis. • Implemented MRCNN, YOLO & performed Transfer Learning using MSCOCO for segmentation with 94% IOU. • Obtained 97.8% accuracy on point cloud clustering using DBSCAN, MeanShift, K-Means and PointNet

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published