Looq AI, Trimble and EagleView use AI for mapping.
As technologies such as artificial intelligence (AI) and machine learning (ML) continue to impact virtually every aspect of daily life, the history-rich field of surveying is proving no exception. Time-tested methods of collecting geospatial data for infrastructure and other engineering projects are being turned upside down by new products that lean heavily on AI and ML. Three recent developments provide a glimpse into the intense activity in new survey data collection and extraction.
New Camera Technology
Looq AI, a California-based company established in 2021, recently introduced a new camera-based system that enables survey-grade capture of 3D data instead of LiDAR scanning and other technologies. The system uses AI-enabled workflows to simplify the process of generating digital twins.
The company’s flagship “Q” product is a handheld, four-camera unit with integrated survey-grade GPS and an AI processor. A mobile phone connected to the back of the unit provides the user interface. The system allows field personnel to walk or drive around a project site and seamlessly capture massive amounts of field data in minutes. After the captured data is uploaded to the cloud, a proprietary algorithm constructs a geo-referenced digital twin with subcentimeter accuracy, according to Dominique Meyer, Looq AI’s cofounder and CEO.
The Looq system offers unique potential to industries such as electric utilities that design, build and operate large-scale systems. With the growth of electric vehicles (EVs) sales and renewable energy usage, electric utility companies are continually evaluating their resources for potential upgrades and new facilities to generate and distribute energy.
“The grid is being shifted to support new uses of electricity, putting a lot of pressure on those resources,” said Meyer. “Looq is focused on supporting this transition. AI has been one of the key enablers to make [the technology] accessible.”
Aquawolf, LLC, an engineering firm supporting the utility industry, has found the Looq AI technology helpful in capturing large datasets quickly. “Their unique hardware and groundbreaking back-end processing has enabled our design teams to capture highly detailed and accurate field conditions in the absence of existing survey data,” said Blake Darling, Aquawolf’s director of operations. “During all phases of design, the platform has enabled us to fill holes in our existing survey basemaps quickly and reliably with rich data, which keeps design moving forward while we wait for traditional land survey.”
More Intelligent Data Extraction
Colorado-based Trimble, a key player in surveying and mapping solutions since 1978, has been incorporating AI technology into its established products to enhance capabilities in areas such as point cloud classification and feature extraction. Both Trimble Business Center (TBC), an office suite that manages geospatial data from multiple sources, and eCognition, a platform for creating custom geospatial analysis solutions, have incorporated AI and deep learning into their recent releases.
In TBC, the new technology has aided the process of converting raw data into usable information with minimal involvement by the end user. “We want a surveyor to be able to automatically convert the data that is acquired with any sensor into actionable information,” said Khrystyna Bezborodova, Trimble product manager.
TBC’s automated classification and feature extraction use a combination of AI techniques, including 2D and 3D deep learning. Classification capabilities are based on a pretrained model that recognizes a broad range of geographic features in point clouds. Feature extraction commands generate geometric attributes for each asset, such as powerlines, manhole covers, trees, pavement markings, poles and signs. For project-specific needs, a new tool for training 3D deep learning models enables users to customize feature extraction workflows to recognize unique features.
In addition to aiding map development, the new capabilities have improved various analysis processes, such as determining stockpile volumes, according to Bezborodova. With point clouds generated from aerial- or ground-based techniques, TBC can automatically extract the boundaries and calculate volumes.
Pavement inspection is also making use of the new technology. Again by using AI technology, TBC can help identify cracks and other distresses, then perform analysis to help users understand the extracted data and make more informed decisions on asset management.
With eCognition, users can further leverage AI technology with custom solutions. The software includes a toolkit to automate feature extraction workflows and additional customization avenues through an integrated Python APl.
For TBC, eCognition and other mapping solutions, the need to combine data from multiple sources will remain key for end users, according to Thomas Widmer, senior product manager for Trimble Photogrammetry. “It’s not that extraction is only happening on points or only happening on imagery, but [the products] have the capability to merge solutions from different content,” he said.
Merging Imagery and Public Safety Solutions
In another recent development, New York-based EagleView, a provider of aerial imagery, software and analytics, has teamed with RapidSOS, a developer of intelligent safety solutions, to integrate EagleView’s high-resolution orthogonal imagery into RapidSOS Premium, a platform that links over 500 million devices to first responder and 911 agencies.
EagleView’s proprietary camera systems capture detailed images with higher resolution and spatial accuracy than standard satellite images, according to the company’s literature. The images also include a date stamp, providing context in emergency situations. In a typical image, each pixel represents 0.75 square inches on the ground. Nearly 20 years of historical imagery is combined with more than 9.5 million miles flown each year to help identify property changes. A computer vision system features deep machine learning algorithms to extract data from current and historical imagery at a scale and speed that far exceeds traditional processes.
RapidSOS Premium aids the 911 response system by consolidating key data such as real-time location, local GIS data and caller profiles into a single, comprehensive mapping solution. The integration with EagleView will enable RapidSOS users to access high-resolution aerial imagery directly in their workflow, providing additional insights into emergency locations and streamlining decision-making for telecommunicators and field responders.
“By integrating EagleView imagery with RapidSOS Premium, we help public safety professionals respond in the most accurate and efficient way to citizens in distress,” said Joe Oddi, Director of Partner Strategies at EagleView.
“Through this alliance with EagleView, telecommunicators can provide more accurate intelligence and directions to support their field responders,” said Karin Marquez, chief public safety brand officer at RapidSOS. “With high-resolution aerial imagery in RapidSOS, public safety officials can make faster, smarter and safer decisions to aid those in the field.”
Future Mapping
These are just a sampling of recent developments in surveying and mapping technology. Numerous other developments are underway to make mapping more accurate, seamless and real time for both AEC professionals and the general public. The increased accuracy may blur lines between GIS and engineering data, enabling more collaboration among different disciplines. Survey data that once found only limited use after project construction will remain viable as part of digital twins and building information modeling (BIM) systems, extending data use from initial project stages into post-construction operations and maintenance.
Future mapping applications may include immersive technologies such as virtual reality and augmented reality, which are already becoming more commonplace in design processes. Applications that once were only accessible on office computers will become more available on mobile devices. And technologies such as AI and ML are destined to play a key role in all new solutions.