Major auto manufacturers have already released, or are soon to release self-driving cars, and for the vehicles to operate safely, purpose-built, accurate, and highly detailed map-based data sets are essential.

Operating self-driving vehicles safely will require these cars to have purpose-built, map-based data sets that contain detailed mapping information with absolute, precise accuracy than the results from current geospatial positioning system resources. To address this need, the Sanborn Map Company has developed proprietary HD mapping technology that created high-precision, 3D maps of California's Santa Clara area, specifically for use in autonomous vehicle models.

Bentley's ContextCapture application enabled Sanborn to create 3D engineering-ready reality meshes of the urban environment through Sanborn's aerial imagery without needing to utilize 3D modelers to create a model of each building, which could have taken up to six months. The real-world 3D data captured in a 3D reality mesh, overlaid with Sanborn HD maps for the Santa Clara area, is being used by automotive partners to simulate their autonomous driving solutions, showcase how their cars will handle different driving scenarios, and help them make informed decisions.

Takeaways:
– Learn how Reality Modeling is used for autonomous driving
– Learn how ContextCapture Center was used to generate reality meshes for the 3D maps
– Building a self-driving technology which can understand the nuances of the world and drive in all the scenarios is a hard problem
– Having an accurate 3D basemap is crucial for driving in bad weather conditions (e.g., snow), when there are no lane markings, complex urban intersections, construction zones, etc
– Creating/updating HD Maps of a city by the sensors on cars could have the side benefit of creating/updating the digital infrastructure (reality model) of the city's road infrastucture