Real-time Analytics
Caltrans needed to inspect and monitor roughly 1,200 active transportation projects more frequently and safely, but manual site visits were slow, inconsistent, and risky for staff working around traffic, bridges, and remote infrastructure.
Reported impact
Caltrans reduced inspection-processing cycles from days or weeks to hours, enabled operators to manage drones from command centers up to 100 miles away, improved worker safety by reducing dangerous site visits, and created a scalable model for persistent infrastructure monitoring.
Caltrans modernized inspections through its SOAR program by deploying autonomous, dock-based Skydio drones equipped with NVIDIA Jetson Orin for edge AI and remote operations. The drones capture aerial imagery in the field, then Caltrans processes the data in a hybrid architecture that spans edge systems and on-premises GPU workstations. Esri ArcGIS provides the NVIDIA-accelerated GIS and geospatial analytics layer, while Trimble uses NVIDIA Omniverse and GPU acceleration libraries to create digital-twin views and 3D project coordination workflows. NVIDIA CUDA and RTX-based acceleration handle photogrammetry, inference, and high-resolution model generation quickly enough for routine operations. This end-to-end stack turns raw flight data into engineering-grade maps, 3D models, and operational insight without requiring staff to physically traverse every inspection site.
Skydio drones collect aerial imagery and site data, Jetson Orin handles edge processing during remote operations, captured data is processed on edge and on-prem GPU systems, ArcGIS and Trimble convert it into geospatial analytics and digital-twin views, and Caltrans teams review the results from command centers instead of hazardous field positions.
Caltrans launched the Scalable Operations and Advanced Remote Technologies (SOAR) program to deploy autonomous Skydio drones equipped with NVIDIA Jetson Orin.
Esri ArcGIS serves as the NVIDIA-accelerated GIS platform, powering GPU-accelerated deep learning and geospatial analytics.
In a hybrid deployment model, captured aerial data is processed both at the edge and on on-premises GPU workstations.
Processes that once took days or weeks are now completed in hours.
Trimble leverages NVIDIA Omniverse™ and GPU acceleration libraries to deliver digital-twin capabilities.
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
NVIDIA Sample Jun 29, 2026 Kaohsiung City Government Kaohsiung needed citywide situational awareness across transportation, flooding, disaster response, and public-safety workflows, but departmental silos and traditional computer vision left agencies unable to interpret… 30,000 diverse smart city camera streams integrated NVIDIA Cosmos · NVIDIA DGX · NVIDIA Metropolis Blueprint for Video Search and Summarization NVIDIA Sample Mar 4, 2026 Israel Meteorological Service The Israel Meteorological Service needed to produce faster and more localized weather forecasts for floods, fire seasons, and other severe conditions, but conventional numerical weather prediction required large CPU… 2.5-kilometer resolution forecasts NVIDIA Earth-2 · NVIDIA Earth-2 Corrector Diffusion (CorrDiff) · NVIDIA PhysicsNeMo NVIDIA Sample Archived · Jan 28, 2026 City of Raleigh Raleigh needed to improve congestion management, mobility, and street safety as the city grew, but manual traffic counts and disconnected sensor repositories made it hard to visualize conditions, respond to incidents… $9.7 million annually in time and fuel costs Esri ArcGIS · Microsoft Azure · NVIDIA Blueprint for video search and summarization (VSS) 