RIEGL and Mach9 Discuss The Impact of AI Feature Extraction on Mobile Mapping Workflows

A recent joint webinar from RIEGL and Mach9 discussed advancements in mobile mapping in terms of hardware and software. Viewers were also given a preview of Mach9’s soon-to-be-released Digital Surveyor software for automating geospatial production and feature extraction from mobile mapping data.

The future of mobile mapping

Mobile mapping advancements can be measured in both improvements in the field and the office. New systems that use a different approach to capture data in the field and hardware updates are examples of collection advancements. Module additions and the integration of AI algorithms are examples of software advancements that are used in the office.

Advancement  can also mean taking a step back, for example, to lower the barrier of entry for mobile mapping. Lower-cost lidar systems are now available for mapping-grade projects aimed at mostly GIS feature extraction of road assets that typically have lower accuracy requirements than survey-grade projects. Such lower-accuracy projects can open doors for lower-cost lidar systems to be deployed. Examples of such lower-cost mobile laser scanning systems are RIEGL’s VMY-1 and VMY-2. These are available in multiple configurations for different projects, including hybrid UAV configurations enabling two-way integration. 

Ladybug6 integration

Because cameras do a better job at identifying assets than just laser data alone, a camera is crucial for asset extraction. As the market is pushing for higher MP cameras to be integrated with lidar systems, RIEGL has made several updates to its camera systems that are integrated with its mobile laser scanning systems, including the integration of the Ladybug6 panorama camera and the RIEGL 24 MP camera. The Ladybug6 is a 72 MP panorama camera that works with RIEGL's both VMY, VMQ-1HA, and VMX-2HA systems.

RiPROCESS is RIEGL’s software package for kinematic lidar data processing and is designed for managing, processing, analyzing, and visualizing data acquired with different laser scanning systems. A new feature in RiPROCESS enables publishing precisely stitched high-resolution orthophotos of the pavement surface, resulting in rectangular tiles with a user-defined tile size.

Extracting features with Mach9’s Digital Surveyor

After capturing mobile mapping data with RIEGL’s RiACQUIRE software, and processing and editing it with RiPROCESS, Mach9’s Digital Surveyor performs the feature extraction. Although data is collected faster than ever before, data processing and extraction have taken a lot of time. Mach9 wants to help partners make extraction nearly as fast as data collection.

Mach9’s Digital Surveyor software consists of three major components: 

• An AI feature extraction tool for automated asset extraction from 3D data;
• QA/QC tools to validate and edit AI-generated workflows;
• Workflow integration with existing CAD and GIS tools.

The demo included a dataset containing 7,590 assets that were automatically extracted in 11 hours and 42 minutes. These were collected during a 108 miles mobile mapping trajectory. The 3D assets were extracted automatically by the software, after which quality assessment takes place. The imagery shows screencaps from the live demo, showing a 3D point cloud, extracted features, and related camera imagery. Using opacity, it is possible to blend the point cloud data with the camera data to get a better view of the scene.

Next, the resulting AI-generated maps can be delivered into CAD and GIS software. Mach9 provides customizable leveling and schemas to do this. A menu inside the software provides different export data formats for downstream use. The list of features that Digital Surveyor can extract includes light poles, utility poles, pedestrian crossing signals, traffic lights, pedestals, signage, sign structures, cabinets, hydrants, ballards, manholes, and a growing list of linear and object features.