Description: Urbanomy is OpalAI's AI-powered multimodal mapping platform that converts LiDAR, panoramic imagery, 4K video, GPS/IMU, crash data, and aerial imagery into a unified digital twin of urban streets.

Objective: Automate road-asset assessment, ADA compliance checks, safety analysis, and multimodal planning.

Pilot Area:
Los Angeles (Koreatown, Hancock Park, Larchmont, East Hollywood) — 7.09 sq. mi, 156.8 road miles.

Partner Agencies:
City of Los Angeles, UCLA, LARIAIC.

Technical Approach:

• Multimodal Mapping: LiDAR, 360° imagery, and 4K video captured >150 miles of roadway.
• Automated Asset Extraction: VLM + LiDAR identified sidewalks, ramps, bike lanes, signage, pavement, and transit features.
• Aerial–Ground Fusion: Combined ground scans with Aerial/ Satellite imagery for better intersection and network coverage.
• Condition Assessment: AI models to score condition of road assets based on safety and quality guidelines from experts.
• LTS Modeling: Developed enhanced Level of Traffic Stress for sidewalks, bike lanes, and crosswalks.
• Cloud Processing: Scalable pipelines processed millions of images and LiDAR points.

Challenges & Opportunities:

• Occlusions & Gaps: Trees, cars, and narrow streets reduced visibility → use sidewalk robots + aerial coverage.
• Large Data Volume: ~40 GB/mile → opportunity for auto-scaling, containerized pipelines.
• Labeling Burden: High manual labeling → adopt active learning + semi-automated annotation.
• Geometric Precision Needs: VLM limitations → expand LiDAR-based ADA geometry measurement.
• Compute Constraints: Cloud Processing Limitations → optimize workflows + hybrid compute strategies.