GeoAI for Large-Scale Image Analysis and Machine Vision: Recent Progress of Artificial Intelligence in Geography

Abstract

GeoAI, or geospatial artificial intelligence, has become a trending topic and the frontier for spatial analytics in Geography. Although much progress has been made in exploring the integration of AI and Geography, there is yet no clear definition of GeoAI, its scope of research, or a broad discussion of how it enables new ways of problem solving across social and environmental sciences. This paper provides a comprehensive overview of GeoAI research used in large-scale image analysis, and its methodological foundation, most recent progress in geospatial applications, and comparative advantages over traditional methods. We organize this review of GeoAI research according to different kinds of image or structured data, including satellite and drone images, street views, and geo-scientific data, as well as their applications in a variety of image analysis and machine vision tasks. While different applications tend to use diverse types of data and models, we summarized six major strengths of GeoAI research, including (1) enablement of large-scale analytics; (2) automation; (3) high accuracy; (4) sensitivity in detecting subtle changes; (5) tolerance of noise in data; and (6) rapid technological advancement. As GeoAI remains a rapidly evolving field, we also describe current knowledge gaps and discuss future research directions.

Mass-Balance-Consistent Geological Stock Accounting: A New Approach toward Sustainable Management of Mineral Resources

“…Data acquisition technologies such as satellites and drones that interact the Internet of Things (IoT) facilitate both global mapping of mining land use, and high-resolution mine-site-scale monitoring of production stockpiles and tailings storage facilities. , Such remote and in situ measurements are key to the extractive industry’s Mining 4.0 vision of smart and connected digital transformation. , It is estimated that 95% of EO data have never been accessed, partly due to challenges with managing its volume, variety, veracity, velocity, and the difficulty to extract value (the five Vs) . This indicates that there is a huge potential for Big Earth Data fusion, geospatial artificial intelligence (GeoAI), and cloud-based computing, which together can help improve data accessibility and support investigative approaches also for users with limited knowledge. , Simultaneously, free or relatively inexpensive access to open government servers or proprietary platforms such as Google’s Earth Engine and Microsoft’s Planetary Computer, coupled with geodata modeling environments including the Open Data Cube (ODC) , and advances in data processing and visualization technologies, − facilitate large-area high-resolution geomodeling. − Digital twins , may soon become standard tools for modeling the geological subsurface together with production facilities at mine-site (plant) scale, and may be part of larger models that integrate geological information with urban-scale building- and city information models (BIM/CIM) into regional GeoBIM systems. , Indeed, two decades after the former Vice President of the USA Al Gore outlined his vision of a “Digital Earth”, the UN-led Coalition for Digital Environmental Sustainability has recently declared the development of a “Planetary Digital Twin” a strategic priority for the sustainability transformation. Given the accelerating rate of innovation, we can imagine multidimensional (e.g., 6D = x , y , z + time + scale/resolution + uncertainty) , Digital Earth Science Platforms − that allow us to model historical, monitor ongoing, and simulate future geological and anthropogenic stock changes and material flows through space and time. Multidimensional Geoinformation Management .…”

“…The industry can benefit from access to previously unavailable information through the B2B data trade. This would allow partners to exploit the collective data volume though machine learning (ML), artificial intelligence (AI), ,, and digital laboratories with augmented and virtual reality (AR/VR), , and can inform mineral systems analysis and exploration, , process innovation, and supply chain management . Similarly, transdisciplinary stakeholder collaborations can contribute to joint problem solving. Policy Trends and Best Practice Examples .…”

“…232 , 233 It is estimated that 95% of EO data have never been accessed, partly due to challenges with managing its volume, variety, veracity, velocity, and the difficulty to extract value (the five Vs). 222 This indicates that there is a huge potential for Big Earth Data fusion, 222 geospatial artificial intelligence (GeoAI), 234 and cloud-based computing, which together can help improve data accessibility and support investigative approaches also for users with limited knowledge. 235 , 236 Simultaneously, free or relatively inexpensive access to open government servers 223 or proprietary platforms such as Google’s Earth Engine 237 and Microsoft’s Planetary Computer, 238 coupled with geodata modeling environments including the Open Data Cube (ODC) 236 , 239 and advances in data processing 240 and visualization technologies, 241 − 243 facilitate large-area high-resolution geomodeling.…”

A horizon scan for novel and impactful areas of physical geography research in 2023 and beyond

“…The rapid evolution of AI hardware and software, and what has been termed ‘big data analytics’ , over the past decade provides an opportunity for enhancing physical geography research by expanding enquiry, diversifying methodologies, and facilitating rapid information extraction from large (sometimes also long time-series) geo-spatial and other datasets. Notably, GeoAI (Li and Hsu, 2022) or Earth AI (Sun et al, 2022) enable researchers to discern patterns and trends in vast volumes of spatial data from sources that can include remote sensing data, climate models, and ecological data, which would be very difficult or impossible through conventional methods. Moreover, these technologies represent an opportunity to bridge quantitative and qualitative approaches in physical geography, human geography, and Geographical Information Science (Ferreira and Vale, 2022).…”

Real-time GeoAI for high-resolution mapping and segmentation of arctic permafrost features

“…Achieving this goal requires new approaches that can perform automated mining from Arctic big data. It is exciting that the Arctic community has started to embrace GeoAI [11,12] and big data to support Arctic research, from predicting Arctic sea ice concentration [1], to finding marine mammals on ice [15], creating Arctic land cover maps [18], and automated mapping of permafrost features [2]. Pioneering research in performing automated characterization of Arctic permafrost features has also been reported in the literature.…”