Advanced Search   
  Look in
       Title     Description
  Include
       Inactive Projects
 
3D Vision and Intelligent Systems Group (3DVIS)
Head: Daniel Huber
Contact: Daniel Huber
Mailing address:
Carnegie Mellon University
Robotics Institute
5000 Forbes Avenue
Pittsburgh, PA 15213

Location:
NSH A501 and A502
412-268-7931
This page last updated - June 2010.
Current Projects
 
Automated Floor Plan Modeling
This project is working to estimate 2D floor plans from sensed 3D data, and to establish criteria for evaluating the accuracy of automated floor plan modeling algorithms.
Automated Reverse Engineering of Buildings
The goal of this project is to use data from 3D sensors to automatically reconstruct compact, accurate, and semantically rich models of building interiors.
Context-based Recognition of Building Components
In this project, we are investigating ways to leverage spatial context for the recognition of core building components, such as walls, floors, ceilings, doors, and doorways for the purpose of modeling interiors using 3D sensor data.
Detailed Wall Modeling in Cluttered Environments
The goal of this project is to develop methods to accurately model wall surfaces even when they are partially occluded and contain numerous openings, such as windows and doorways.
E57 Standard for 3D Imaging System Data Exchange
The goal of this project is to develop a vendor-neutral data exchange format for data produced by 3D imaging systems, such as laser scanners.
LIDAR and Vision Sensor Fusion for Autonomous Vehicle Navigation
The goal of this project is to investigate methods for combining laser range sensors (i.e., LIDARs) with visual sensors (i.e., video cameras) to improve the capabilities of autonomous vehicles.
Moving Object Detection, Modeling, and Tracking
The goal of this research is to better understand how vision and 3D LIDAR data be combined for detecting and tracking moving objects.
Optimal LIDAR Sensor Configuration
This project is developing a framework that allows objective comparison between alternative LIDAR configurations.
Quality Assessment of As-built Building Information Models using Deviation Analysis
The goal of this project is to develop a method for conducting quality assessment (QA) of as-built building information models (BIMs) that utilizes patterns in the differences between the data within and between steps in the as-built BIM creation process to identify potential errors.
Real-time Lane Tracking in Urban Environments
The purpose of this project is to develop methods for the real-time detection and tracking of lanes and intersections in urban scenarios in order to support road following by an autonomous vehicle in GPS-denied situations.
Representation of As-built BIMs
This project is investigating how the imperfections of sensed 3D data can be represented within the context of the BIM framework, which was originally designed to handle only perfect data from CAD systems.
Terrain Estimation using Space Carving Kernels
This project uses information about the ray extending from the sensor to the sensed surface be used to improve terrain estimation in unstructured environments.
Tightly Integrated Stereo and LIDAR
The goal of this project is to use sparse, but accurate 3D data from LIDAR to improve the estimation of dense stereo algorithms in terms of accuracy and speed.
Transforming Surface Representations to Volumetric Representations
This project’s goal is to transform the surface-based representations that are naturally derived from sensed data into volumetric representations needed by CAD and BIM.
Vehicle Localization in Naturally Varying Environments
The purpose of this project is to develop methods for place matching that are invariant to short- and long-term environmental variations in support of autonomous vehicle localization in GPS-denied situations.