← Back to Index

Segmentation Gallery

Semantic segmentation showing vegetation detection for GVI calculation

Vegetation
Sky
Building
Road/Ground
Person
Vehicle

How GVI is Calculated

The Green View Index (GVI) measures the percentage of vegetation visible in a street-level image. A deep learning model (semantic segmentation) identifies different objects in each image. The green-colored pixels represent detected vegetation (trees, shrubs, grass). GVI = (green pixels) / (total pixels) × 100%

Al Karama's mean GVI is only 3.5%, indicating very low street-level vegetation visibility.

🌿 High GVI Examples (50-65% vegetation)
High GVI segmentation
65.0% GVI
Highest vegetation visibility in Al Karama
View original on Mapillary →
High GVI segmentation
63.5% GVI
Dense tree canopy coverage
View original on Mapillary →
High GVI segmentation
61.8% GVI
Street with mature trees
View original on Mapillary →
High GVI segmentation
60.9% GVI
Green corridor area
View original on Mapillary →
High GVI segmentation
59.7% GVI
Tree-lined street
View original on Mapillary →
🏜️ Low GVI Examples (0% vegetation)
Low GVI segmentation
0.0% GVI
No visible vegetation
View original on Mapillary →
Low GVI segmentation
0.0% GVI
Urban canyon - buildings only
View original on Mapillary →
Low GVI segmentation
0.0% GVI
Concrete dominated view
View original on Mapillary →
Low GVI segmentation
0.0% GVI
No greenery present
View original on Mapillary →
Low GVI segmentation
0.0% GVI
Built environment only
View original on Mapillary →

Segmentation performed using ZenSVI with Mapillary/Cityscapes trained models

← Back to Index