Building facade reconstruction by fusing terrestrial laser points and images

Area of Science:

• Geomatics Engineering
• Computer Vision
• 3D Reconstruction

Background:

• 3D feature extraction benefits from complementary laser and optical data.
• Integrating diverse data sources improves reliability and automation.
• Building facade reconstruction is crucial for urban modeling.

Purpose of the Study:

• To develop a semiautomatic approach for 3D building facade reconstruction.
• To efficiently combine terrestrial laser point clouds and close-range images.
• To enhance the accuracy and automation of 3D feature extraction.

Main Methods:

• Utilizing planar features from laser data to establish facade structure.
• Extracting strong image lines using Canny and Hough transformations.
• Comparing image-derived edges with existing model edges for refinement.
• Selecting and applying textures based on optimal visibility and image orientation.

Main Results:

• Successfully reconstructed building facades by integrating laser and image data.
• Addressed challenges in referencing laser points with multiple images.
• Achieved automated texturing with optimal visibility.
• Demonstrated a semiautomatic approach for improved 3D feature extraction.

Conclusions:

• The combined use of laser and optical data offers a robust solution for 3D facade reconstruction.
• The developed method enhances automation and accuracy in 3D modeling.
• Further research is needed to address remaining limitations and expand capabilities.