Method and system for evaluating urban cycling traffic resilience based on street view perception and machine learning
By using a street view perception and machine learning approach, tree canopies, bike lanes, sky, and shaded areas in street view images are extracted, and lateral misalignment and sky opening are calculated. This solves the problem of insufficient identification of bike lane sun exposure risk under high green visibility in existing technologies, and achieves a more accurate assessment of the thermal resilience of cycling traffic.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- TONGJI UNIV
- Filing Date
- 2026-06-15
- Publication Date
- 2026-07-17
AI Technical Summary
Existing urban cycling environment assessment methods fail to accurately identify the risk of continuous sun exposure on the center track of bike lanes under high green visibility conditions, resulting in inconsistencies between thermal resilience assessment results and the actual risks to cyclists.
Using a street view perception and machine learning approach, the system extracts tree canopy, bike path, sky and shade areas through an image semantic segmentation model, calculates lateral misalignment, sky opening and shade coverage, generates reverse thermal toughness risk values, and identifies the sun exposure risk of cycling routes.
It improves the accuracy of urban cycling traffic thermal resilience assessment, can identify the sun exposure risk of the center track of the bike lane under high green visibility, provides interpretable image analysis basis, and supports traffic environment optimization and cycling access improvement.
Smart Images
Figure CN122415599A_ABST