Urban Forestry Mapping

Street-View Tree Census

This project explores how a city-scale tree census can be built from street-view imagery instead of sending survey teams to count every trunk manually.

What This Shows

We used a custom-trained model to detect tree trunks in street-view images and then map those detections across the surveyed region.

The result is a fast, repeatable way to estimate urban tree inventory at a city scale. In the limited areas where we ran ground-truth validation, the approach performed with strong accuracy.

Why it matters A workflow like this can help cities update tree counts, prioritize field work, and monitor urban canopy infrastructure without starting from scratch each time.

Layers

Base Map

Selection

Click a mapped tree to inspect its location and detection metadata.
Contributors Malhar Bhise Utkarsh Agarwal Bhavi
Mentored by Dr Anupam Sobti.

This public demo focuses on the mapped outputs. Panorama rendering and model internals are intentionally omitted here to keep the site lightweight and easy to publish.