Geology topics

Sizhe Wang

Publications and source records attributed to Sizhe Wang.

4 recordsLinked to original sources

USGS2024

Segment anything model can not segment anything: Assessing AI foundation model's generalizability in permafrost mapping

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.

Remote Sensing↗

USGS2023

Assessment of a new GeoAI foundation model for floodinundation mapping

Vision foundation models are a new frontier in GeoAI research because of their potential to enable powerful image analysis by analyzing and extracting important image features from vast amounts of geospatial data. This paper evaluates the performance of the first-of-its-kind geospatial foundation model, IBM-NASA’s Prithvi, to support a crucial geospatial analysis task: flood inundation mapping. This model is compared with popular convolutional neural networks and vision transformer-based architectures regarding mapping accuracy for flooded areas. A benchmark dataset, Sen1Floods11, is used in the experiments, and the models' predictability, generalizability, and transferability are evaluated based on both validation datasets and datasets completely unseen by the model. Results show the impressive transferability of the Prithvi model, highlighting its performance advantages in segmenting flooded areas in previously unseen regions. The findings also suggest areas for improvement for the Prithvi model in adopting multi-scale representation learning, developing more end-to-end pipelines for high-level image analysis tasks, and offering more flexibility in allowable input data bands.

Conference Paper↗

USGS2022

GeoImageNet: A multi-source natural feature benchmark dataset for GeoAI and supervised machine learning

The field of GeoAI or Geospatial Artificial Intelligence has undergone rapid development since 2017. It has been widely applied to address environmental and social science problems, from understanding climate change to tracking the spread of infectious disease. A foundational task in advancing GeoAI research is the creation of open, benchmark datasets to train and evaluate the performance of GeoAI models. While a number of datasets have been published, very few have centered on the natural terrain and its landforms. To bridge this gulf, this paper introduces a first-of-its-kind benchmark dataset, GeoImageNet, which supports natural feature detection in a supervised machine-learning paradigm. A distinctive feature of this dataset is the fusion of multi-source data, including both remote sensing imagery and DEM in depicting spatial objects of interest. This multi-source dataset allows a GeoAI model to extract rich spatio-contextual information to gain stronger confidence in high-precision object detection and recognition. The image dataset is tested with a multi-source GeoAI extension against two well-known object detection models, Faster-RCNN and RetinaNet. The results demonstrate the robustness of the dataset in aiding GeoAI models to achieve convergence and the superiority of multi-source data in yielding much higher prediction accuracy than the commonly used single data source.

GeoInformatica↗

USGS2020

GeoNat v1.0: A dataset for natural feature mapping with artificial intelligence and supervised learning

Machine learning allows “the machine” to deduce the complex and sometimes unrecognized rules governing spatial systems, particularly topographic mapping, by exposing it to the end product. Often, the obstacle to this approach is the acquisition of many good and labeled training examples of the desired result. Such is the case with most types of natural features. To address such limitations, this research introduces GeoNat v1.0, a natural feature dataset, used to support artificial intelligence‐based mapping and automated detection of natural features under a supervised learning paradigm. The dataset was created by randomly selecting points from the U.S. Geological Survey’s Geographic Names Information System and includes approximately 200 examples each of 10 classes of natural features. Resulting data were tested in an object‐detection problem using a region‐based convolutional neural network. The object‐detection tests resulted in a 62% mean average precision as baseline results. Major challenges in developing training data in the geospatial domain, such as scale and geographical representativeness, are addressed in this article. We hope that the resulting dataset will be useful for a variety of applications and shed light on training data collection and labeling in the geospatial artificial intelligence domain.

Transactions in GIS↗