Creation, Transformation, and Orientation Adjustment of a Building Facade Model for Feature Segmentation: Transforming 3D Building Point Cloud Models into 2D Georeferenced Feature Overlays

Abstract

This project involves the creation of 3-dimensional models of building facades from ground-derived point clouds and extracting features from the facades as encoded overlays. These features include balconies, window wells, overhangs, or other exterior fixtures. Point data of a building facade, acquired from high-resolution terrestrial scans, are transformed in space so that a normal to the facade surface is in the vertical direction. The point cloud is then converted into a gridded model, which can be investigated in like manner to a digital terrain model through breakline, elevation difference, and slope analysis. A backplane of constant z values is created underneath the facade to act as a bare earth surface. The orientation of the backplane in space is carefully adjusted to be parallel to the facades exterior, providing more accurate feature extraction through difference analysis. Raster overlays are created as color-mapped classes or continuous values of breakline, elevation difference, or slope that capture the form of facade feature objects. To complete the attribution of facade features, georeferencing information from the original building model is associated with each overlay product. This allows for the display of the overlays in a georeferenced visualization environment such as a Geographic Information System (GIS) or Google Earth.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 2019
Accession Number
AD1089177

Entities

People

  • Philip Devine
  • S. B. Blundell

Tags

Communities of Interest

  • Air Platforms
  • Materials and Manufacturing Processes
  • Space

DTIC Thesaurus Topics

  • Computer Vision
  • Data Analysis
  • Data Visualization
  • Detection
  • Earth Models
  • Feature Extraction
  • Geographic Information Systems
  • Geometry
  • Global Positioning Systems
  • Graphical User Interface
  • High Resolution
  • Information Systems
  • Models
  • Point Clouds
  • Terrain Models
  • Three Dimensional
  • Visualizations

Readers

  • Computer Vision.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • Space