An algorithm for ice/water detection of ice less than 30 cm in thickness (thin ice) using dual-polarized synthetic aperture radar (SAR) and passive microwave data has been developed based on insights from previous work. The frequency chosen from those available was in the C-band. For the radiometer layer ice concentrations, land contaminated pixels were removed. For the SAR layers, a novel approach to their use in ice detection was developed using the coefficient of variation (COV), defined as the ratio of the standard deviation of the normalized radar cross section, σ0, to the mean σ0. The COV reduced the complicated signatures of breaking waves and thin ice to bright ice pixels on a dark water background. The co-pol and cross-pol COV layers and the radiometer layer were merged according to physics-based rules, except in regions adjacent to land or the ice edge where only the SAR layers were valid. The detection algorithm was applied to 233 Sentinel 1a and 1b dual-polarization images of the freeze-up from 2015, 2016, 2018, and 2019 in the Beaufort and Chukchi Seas. The synergistic ice retrievals showed improved agreement with manual operational analyses in the marginal ice zone over other automated operational products.

Semantic segmentation is a powerful method to facilitate visual scene understanding. Each pixel is assigned a label according to a pre-defined list of object classes and semantic entities. This becomes very useful as a means to summarize large scale overhead imagery. In this paper we present our work on semantic segmentation with applications to overhead imagery. We propose an algorithm that builds and extends upon the DeepLab framework to be able to refine and resolve small objects (relative to the image size) such as vehicles. We have also investigated sensor adaptation as a means to augment available training data to be able to reduce some of the shortcomings of neural networks when deployed in new environments and to new sensors. We report results on several datasets and compare performance with other state-of-the-art architectures.

Recent advances in Generative Adversarial Learning allow for new modalities of image super-resolution by learning low to high resolution mappings. In this paper we present our work using Generative Adversarial Networks (GANs) with applications to overhead and satellite imagery. We have experimented with several state-of-the-art architectures. We propose a GAN-based architecture using densely connected convolutional neural networks (DenseNets) to be able to super-resolve overhead imagery with a factor of up to 8x. We have also investigated resolution limits of these networks. We report results on several publicly available datasets, including SpaceNet data and IARPA Multi-View Stereo Challenge, and compare performance with other state-of-the-art architectures.

This paper presents a radar sensor package specifically developed for wide-coverage sounding and imaging of polar ice sheets from a variety of aircraft. Our instruments address the need for a reliable remote sensing solution well-suited for extensive surveys at low and high altitudes and capable of making measurements with fine spatial and temporal resolution. The sensor package that we are presenting consists of four primary instruments and ancillary systems with all the associated antennas integrated into the aircraft to maintain aerodynamic performance. The instruments operate simultaneously over different frequency bands within the 160 MHz-18 GHz range. The sensor package has allowed us to sound the most challenging areas of the polar ice sheets, ice sheet margins, and outlet glaciers; to map near-surface internal layers with fine resolution; and to detect the snow-air and snow-ice interfaces of snow cover over sea ice to generate estimates of snow thickness. In this paper, we provide a succinct description of each radar and associated antenna structures and present sample results to document their performance. We also give a brief overview of our field measurement programs and demonstrate the unique capability of the sensor package to perform multifrequency coincidental measurements from a single airborne platform. Finally, we illustrate the relevance of using multispectral radar data as a tool to characterize the entire ice column and to reveal important subglacial features.

This work focuses on automating the task of estimating Polar ice thickness from airborne radar data acquired over Greenland and Antarctica. This process involves the identification and accurate selection of the ice sheet's surface location and interface between the ice sheet and the underlying bedrock for each measurement. Identifying the surface and bedrock locations in the radar imagery enables the computation of ice sheet thickness, which is important for the study of ice sheets, their volume, and how they may contribute to global climate change. The time-consuming manual approach requires sparse hand-selection of surface and bedrock interfaces by several human experts, and interpolating between the selections to save time.

This paper focuses on automating the task of estimating Polar ice thickness from airborne radar data acquired over Greenland and Antarctica. This process involves the identification and accurate selection of the ice sheet's surface location and interface between the ice sheet and the underlying bedrock for each measurement. Identifying the surface and bedrock locations in the radar imagery enables the computation of ice sheet thickness, which is important for the study of ice sheets, their volume, and how they may contribute to global climate change. The time-consuming manual approach requires sparse hand-selection of surface and bedrock interfaces by several human experts, and interpolating between the selections to save time. Two primary methods have been studied in this paper, namely, edge-based and active contour. Results are evaluated and presented using the metrics of time requirements and accuracy. Automated ice thickness estimation results from 2006 and 2007 Greenland field campaigns illustrate that the edge-based approach offers faster processing (seconds compared to minutes), but suffers from a lack of continuity and smoothness aspects that active contours provide. The active contour approach is more accurate when compared to ground truth selections provided by human experts, and has proven to be more robust to image artifacts. It is shown that both techniques offer advantages which could be integrated to yield a more effective system.

This paper is focused on the effects of sharing knowledge and collaboration of multiple heterogeneous, intelligent agents (hardware or software) which work together to learn a task. As each agent employs a different machine learning technique, the system consists of multiple knowledge sources and their respective heterogeneous knowledge representations. Collaboration between agents involves sharing knowledge to both speed up team learning, as well as refine the team's overall performance and group behavior. Experiments have been performed that vary the team composition in terms of machine learning algorithms, learning strategies employed by the agents, and sharing frequency for a predator-prey cooperative pursuit task. For lifelong learning, heterogeneous learning teams were more successful than homogeneous learning counterparts. Interestingly, sharing increased the learning rate, but sharing with higher frequency showed diminishing results. Lastly, knowledge conflicts are reduced over time the more sharing takes place. These results support further investigation of the merits of heterogeneous learning. © 2008 Wiley Periodicals, Inc.