Automated Building Material Detection Using 3D Side Polygons
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automating the identification of materials in geospatial imagery is challenging due to the position and orientation of satellite imagery, which can obscure portions of physical objects, making it difficult to accurately detect and classify building materials without human intervention.
Innovation Solution
The proposed technology generates building side polygons using off-nadir, multispectral images, building footprint data, and elevation data, employing vector analysis to determine visible side faces and associate them with building footprints, enabling automated identification and classification of building materials.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated material identification is implemented using machine learning, then productivity increases and manual review is reduced, but measurement precision deteriorates due to occluded portions of buildings in off-nadir imagery
Solution Approach 1:
The patent transitions from analyzing only 2D image data to incorporating 3D building geometry information. By generating building side polygons that represent three-dimensional structures and combining them with elevation data, the system compensates for occlusions in off-nadir imagery. This dimensional enhancement allows the automated system to infer material properties of obscured building portions, thereby maintaining measurement precision while preserving automated processing efficiency.
Solution Approach 2:
The patent performs preliminary generation of building side polygons and occlusion analysis before material identification. By pre-processing the imagery to identify visible and occluded portions of buildings based on satellite viewing geometry, the system prepares corrected or supplemented data structures that enable accurate automated material detection despite initial occlusions. This preliminary action ensures that the machine learning model receives enhanced input data that accounts for geometric occlusions.
2Area of stationary object
If off-nadir satellite imagery is used, then area coverage increases and more building sides become visible, but measurement precision worsens due to occluded portions of buildings
Solution Approach 1:
The patent segments the building into multiple side polygons, each representing a different facade. By analyzing the satellite viewing angle and building geometry, the system identifies which side polygons are visible and which are occluded in off-nadir imagery. This segmentation allows the system to process only the visible portions for direct material identification while using geometric relationships and elevation data to infer properties of occluded portions, thereby maintaining measurement precision across the entire building despite limited viewing angles.
Solution Approach 2:
The patent introduces building side polygons as an intermediary data structure between the raw satellite imagery and the material identification process. These polygons serve as a geometric model that mediates between the off-nadir viewing geometry and the material detection algorithm. By using these intermediate geometric representations combined with elevation data, the system can accurately determine which building surfaces are visible and infer properties of occluded surfaces, thus maintaining measurement precision while benefiting from the broader coverage of off-nadir imagery.
Data Source
AI summary
A system and method for automatically (without human intervention) identifying a material in an image that comprises a building material for buildings in the image. Building side polygons which may be used to identify building sides in off-nadir imagery are generated. Off-nadir, multispectral images, building footprint data and elevation data for a geographic area are taken as input. Building heights for buildings in the geographic area are determined by clipping the elevation data using the building footprint data and then calculating building heights. A candidate set of polygons representing visible side faces of each building in the images is created from the known building heights, and based on the viewpoint direction, using vector analysis. After culling occluded polygons and polygons too small for analysis, the polygons are associated with a building footprint. Building materials for each building having visible polygons can then be identified.