Artificial Intelligence (AI) is not a new technology, but tools utilizing it have been experiencing a boom over the last year as large language models and image generators have begun to break into the mainstream discourse. As a result, we’ve seen all types of AI receive more attention and funding, including in geospatial space, with the expectation that rapid development will only continue. In this session, learn how different types of AI are already being applied in geospatial applications, and the potential that some of the bleeding-edge tools have to change the way we capture, analyze and utilize geospatial data.Â
AI-Powered Point Cloud Processing Opportunities and Challenges
Artificial intelligence is revolutionizing our daily work routines and is now making its way into the geospatial domain. To stay ahead in this rapidly evolving landscape, it’s crucial to embrace Artificial Intelligence and unlock its full potential for extracting value.
In this talk, we will take a deep dive into the topic of using Artificial Intelligence for point cloud processing. We will explore the fundamental concepts and core principles behind this state-of-the-art technology. We will shed light on the quality metrics, and why they are essential. Choosing inappropriate metrics can skew our results in unwanted directions, for no obvious reasons. We will also uncover the unique advantages of leveraging Artificial Intelligence and machine learning in geospatial applications and explore real-world use cases where Artificial Intelligence has been applied, ranging from forest inventory creation to 3D building reconstructions.
As we move rapidly into a future powered by Artificial Intelligence, it becomes imperative for us to harness its potential in all spheres, including the geospatial realm. By understanding and leveraging Artificial Intelligence for point cloud processing, we can reveal, optimize, and solve existing and new use cases.
Nejc Dougan, FlaiÂ
Environmental Twinning: A New Way to Manage Asset Resilience in the Face of a Changing Climate
Anticipated climate change adaptation costs are projected to heavily impact infrastructure and could result in an average annual investment of $150B – $450B by the year 2050. In this era of unprecedented environmental challenges, the traditional focus on creating a digital twin of infrastructure is not enough. Organizations must think beyond the asset footprint to quantify exposures to environmental threats and develop effective resilience strategies. “Environmental Twinning” emerges as a promising strategy that combines remote-sensing technology, ecosystem science, and data-driven insights to model the environment surrounding infrastructure and capture change over time.
The presentation will explore Environmental Twinning as a novel approach to asset resilience. By leveraging advanced data science and high-performance processing, organizations can scale remotely-sensed data to create temporal and spatial environmental twins. These digital replicas provide a dynamic representation of the changing landscape impacting infrastructure, enabling real-time monitoring, analysis, and predictive modeling.
The presentation aims to inspire participants to rethink their approach to asset management in the face of a changing climate. By embracing environmental twins, organizations can keep a pulse on changing asset vulnerabilities, enhance decision-making processes, and develop proactive strategies to build resilient infrastructure that can withstand the challenges of the future.
Tobias Kraft, Teren, Inc.
AI for GIS: Processes for Digital Twins that can be Realized Today
Over the last year, AI has become an ubiquitous term in all areas of business – it crosses industries, markets, trades, and beyond. The AI market value of nearly 100 billion U.S. dollars is expected to grow twentyfold by 2030, up to nearly two trillion U.S. dollars. It is especially prevalent at GeoWeek in construction, mapping, smart cities applications, mobility, transportation, digital twins and more. Dr. Aaron Morris, CEO of Allvision (creator of the AIGIS software platform) will discuss the key points of delivering AI for GIS-related products and services, including how to ensure quality data, why and how to maintain foundation models, and how to assure a functional feedback loop for the best data results. For GIS managers and engineers, AI can improve project outcome quality while reducing costs of project delivery, saving human-in-the-loop time, and enhancing the overall findings for better decision making. The more you use AI, the better it becomes! Interested in integrating AI into your processes? Aaron will show you how, including case study examples of major metropolitan areas mapping their assets with AI-produced digital twins, including signage, bus stops, parking, and guide rails. There are many things AI can do, but don’t forget what it cannot – Aaron will debunk several processes that haven’t been realized as of yet and what to utilize instead.
Aaron Morris, Allvision
AI/ML-Driven Pavement Crack Detection
Crack detection on pavement has always been a hard and tedious task to complete. Although these defects can cause critical failures on infrastructure, the sheer difficulty of finding, assessing and reporting them lead to an inaccurate representation of reality. This is where Artificial Intelligence and Machine Learning enabled technologies can speed up the process and more accurately report pavement cracks. Â Â
In this joint presentation with Benesch, we will show how traditional on-the-ground methods for listing and reporting are highly inefficient with inspectors having to walk the pavement to take photos while making handwritten notations and taking measurements with a tape or wheel. The advent of drone technology to capture high-resolution orthomosaic aerials has allowed for tracing cracks directly on images, saving time in the digitization process. However, it does not allow practitioners to analyze or classify cracks based on a condition assessment without the accompanying boots-on-the-ground inspection activities. Â Â
We will then show how AI crack detector model is used to detect, classify, report and export the output results. The model was tested on three separate projects, providing us with a reliable analysis of crack data from the project sites while saving time and labor in the field. It also eliminated the time needed for digitizing crack linework data, saving time and money in the current context of labor shortages.
Gen Taurand, Bentley Systems and Bret Tremblay, Alfred Benesch & Associates
