Oakland 3-D Point Cloud Dataset - CVPR 2009 subset

This repository contains labeled 3-D point cloud laser data collected from a moving platform in a urban environment. Data are provided for research purposes.
Please submit questions or comments to Nicolas Vandapel (vandapel@ri.cmu.edu)

This data set was used to produce the results presented in our CVPR 2009 paper [project page]. If you use the data, please reference the following publication:
     Contextual Classification with Functional Max-Margin Markov Networks.
     Daniel Munoz, J. Andrew (Drew) Bagnell, Nicolas Vandapel, and Martial Hebert.
     IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), June, 2009.

The data was collected using Navlab11 equiped with side looking SICK LMS laser scanners and used in push-broom. The data was collected around CMU campus in Oakland, Pittsburgh, PA.

Data are provided in ascii format: x y z label confidence, one point per line, space as separator. Corresponding vrml files (*.wrl) and label counts (*.stats) are also provided. The data set is made of two subset (part2, part3) with each its own local reference frame, where each file contains 100,000 3-D points. The training/validation and testing data was filtered and labeled remapped from 44 into 5 labels [TXT].

	Complete	Training	Validation	Testing
statistics	17 files, 1.6 millions 3-D pts, 44 labels	1 file, 36932 3-D pts, 5 labels	1 file, 91579 3-D pts, 5 labels	15 files, 1.3 millions 3-D pts, 5 labels
Label distribution	[PNG]	[PNG]	[PNG]	[PNG]
Snapshots		[PNG]	[PNG]
Data set	[ZIP, 16 MB]	[ZIP, 2 MB]	[ZIP, 1 MB]	[ZIP, 14 MB]

Prepared through collaborative participation in the Robotics consortium sponsored by the U.S Army Research Laboratory under the Collaborative Technology Alliance Program, Cooperative Agreement DAAD19-01-209912.

revision log