The microtopography associated with ice-wedge polygons governs many aspects of Arctic ecosystem, permafrost, and hydrologic dynamics from local to regional scales owing to the linkages between microtopography and the flow and storage of water, vegetation succession, and permafrost dynamics. Wide-spread ice-wedge degradation is transforming low-centered polygons into high-centered polygons at an alarming rate. Accurate data on spatial distribution of ice-wedge polygons at a pan-Arctic scale are not yet available, despite the availability of sub-meter-scale remote sensing imagery. This is because the necessary spatial detail quickly produces data volumes that hamper both manual and semi-automated mapping approaches across large geographical extents. Accordingly, transforming big imagery into ‘science-ready’ insightful analytics demands novel image-to-assessment pipelines that are fueled by advanced machine learning techniques and high-performance computational resources. In this exploratory study, we tasked a deep-learning driven object instance segmentation method (i.e., the Mask R-CNN) with delineating and classifying ice-wedge polygons in very high spatial resolution aerial orthoimagery. We conducted a systematic experiment to gauge the performances and interoperability of the Mask R-CNN across spatial resolutions (0.15 m to 1 m) and image scene contents (a total of 134 km²) near Nuiqsut, Northern Alaska. The trained Mask R-CNN reported mean average precisions of 0.70 and 0.60 at thresholds of 0.50 and 0.75, respectively. Manual validations showed that approximately 95% of individual ice-wedge polygons were correctly delineated and classified, with an overall classification accuracy of 79%. Our findings show that the Mask R-CNN is a robust method to automatically identify ice-wedge polygons from fine-resolution optical imagery. Overall, this automated imagery-enabled intense mapping approach can provide a foundational framework that may propel future pan-Arctic studies of permafrost thaw, tundra landscape evolution, and the role of high latitudes in the global climate system.

Traditional methods of detecting and mapping utility poles are inefficient and costly because of the demand for visual interpretation with quality data sources or intense field inspection. The advent of deep learning for object detection provides an opportunity for detecting utility poles from side-view optical images. In this study, we proposed using a deep learning-based method for automatically mapping roadside utility poles with crossarms (UPCs) from Google Street View (GSV) images. The method combines the state-of-the-art DL object detection algorithm (i.e., the RetinaNet object detection algorithm) and a modified brute-force-based line-of-bearing (LOB, a LOB stands for the ray towards the location of the target [UPC at here] from the original location of the sensor [GSV mobile platform]) measurement method to estimate the locations of detected roadside UPCs from GSV. Experimental results indicate that: (1) both the average precision (AP) and the overall accuracy (OA) are around 0.78 when the intersection-over-union (IoU) threshold is greater than 0.3, based on the testing of 500 GSV images with a total number of 937 objects; and (2) around 2.6%, 47%, and 79% of estimated locations of utility poles are within 1 m, 5 m, and 10 m buffer zones, respectively, around the referenced locations of utility poles. In general, this study indicates that even in a complex background, most utility poles can be detected with the use of DL, and the LOB measurement method can estimate the locations of most UPCs.

We developed a high-throughput mapping workflow, which centers on deep learning (DL) convolutional neural network (CNN) algorithms on high-performance distributed computing resources, to automatically characterize ice-wedge polygons (IWPs) from sub-meter resolution commercial satellite imagery. We applied a region-based CNN object instance segmentation algorithm, namely the Mask R-CNN, to automatically detect and classify IWPs in North Slope of Alaska. The central goal of our study was to systematically expound the DLCNN model interoperability across varying tundra types (sedge, tussock sedge, and non-tussock sedge) and image scene complexities to refine the understanding of opportunities and challenges for regional-scale mapping applications. We corroborated quantitative error statistics along with detailed visual inspections to gauge the IWP detection accuracies. We found promising model performances (detection accuracies: 89% to 96% and classification accuracies: 94% to 97%) for all candidate image scenes with varying tundra types. The mapping workflow discerned the IWPs by exhibiting low absolute mean relative error (AMRE) values (0.17–0.23). Results further suggest the importance of increasing the variability of training samples when practicing transfer-learning strategy to map IWPs across heterogeneous tundra cover types. Overall, our findings demonstrate the robust performances of IWPs mapping workflow in multiple tundra landscapes.

Retrogressive thaw slumps (RTS) are considered one of the most dynamic permafrost disturbance features in the Arctic. Sub-meter resolution multispectral imagery acquired by very high spatial resolution (VHSR) commercial satellite sensors offer unique capacities in capturing the morphological dynamics of RTSs. The central goal of this study is to develop a deep learning convolutional neural net (CNN) model (a UNet-based workflow) to automatically detect and characterize RTSs from VHSR imagery. We aimed to understand: (1) the optimal combination of input image tile size (array size) and the CNN network input size (resizing factor/spatial resolution) and (2) the interoperability of the trained UNet models across heterogeneous study sites based on a limited set of training samples. Hand annotation of RTS samples, CNN model training and testing, and interoperability analyses were based on two study areas from high-Arctic Canada: (1) Banks Island and (2) Axel Heiberg Island and Ellesmere Island. Our experimental results revealed the potential impact of image tile size and the resizing factor on the detection accuracies of the UNet model. The results from the model transferability analysis elucidate the effects on the UNet model due the variability (e.g., shape, color, and texture) associated with the RTS training samples. Overall, study findings highlight several key factors that we should consider when operationalizing CNN-based RTS mapping over large geographical extents.

This paper assesses trending AI foundation models, especially emerging computer vision foundation models and their performance in natural landscape feature segmentation. While the term foundation model has quickly garnered interest from the geospatial domain, its definition remains vague. Hence, this paper will first introduce AI foundation models and their defining characteristics. Built upon the tremendous success achieved by Large Language Models (LLMs) as the foundation models for language tasks, this paper discusses the challenges of building foundation models for geospatial artificial intelligence (GeoAI) vision tasks. To evaluate the performance of large AI vision models, especially Meta’s Segment Anything Model (SAM), we implemented different instance segmentation pipelines that minimize the changes to SAM to leverage its power as a foundation model. A series of prompt strategies were developed to test SAM’s performance regarding its theoretical upper bound of predictive accuracy, zero-shot performance, and domain adaptability through fine-tuning. The analysis used two permafrost feature datasets, ice-wedge polygons and retrogressive thaw slumps because (1) these landform features are more challenging to segment than man-made features due to their complicated formation mechanisms, diverse forms, and vague boundaries; (2) their presence and changes are important indicators for Arctic warming and climate change. The results show that although promising, SAM still has room for improvement to support AI-augmented terrain mapping. The spatial and domain generalizability of this finding is further validated using a more general dataset EuroCrops for agricultural field mapping. Finally, we discuss future research directions that strengthen SAM’s applicability in challenging geospatial domains.

Increasing availability and advancements of aerial Light Detection and Ranging (LiDAR) data have radically been shifting the way archeological surveys are performed. Unlike optical remote sensing imagery, LiDAR pulses travel through small gaps in dense tree canopies enabling archeologists to discover “hidden” past settlements and anthropogenic landscape features. While LiDAR has been increasingly adopted in archeological studies worldwide, its full potential is still being explored in the United States. Furthermore, while hand-digitizing features in remote-sensing datasets remain a valuable method for archeological surveys, it is often time- and labor intensive. The central objective of this research is to develop a geographic object-based image analysis-driven methodological framework linking low-level features and domain knowledge to automatically extract targets of interest from LiDAR-based digital terrain models (DTMs) and to closely examine the degree of interoperability of knowledge-based rulesets across different study sites focusing on the same semantic class. We apply this framework in southern New England, a geographic region in the northeastern United States where numerous seventeenth-century to early twentieth-century features such as relict charcoal hearths (RCHs), stone walls, and building foundations lie abandoned in densely forested terrain. Focusing on RCHs in this study, our results show promising agreement between manual and automated detection of these features. Overall, we show that the use of LiDAR data augmented with object-based classification workflows provides valuable baseline data for future archeological study and reconstruction of land-use/land-cover change over the past 300 years.