Speakers and topics subject to change.

This presentation deals with global security challenges in the present day, as well as projections for U.S. and allied nations’ responses to ongoing developments throughout the globe. It provides an overview of geographic and trans-national trends, including implications for both government and industry. All of this is viewed with an emphasis on GEOINT matters, in response to evolving resource decisions, tradecraft, technical advances, operations and strategy.

This presentation will address the notion that the future of Geospatial Information Technology belongs to the simple and the quick. The assertion is that we are witnessing a new normal emerging within the Geospatial community whereby many incoming members of this industry increasingly are professionals with backgrounds across multiple disciplines but they are not necessarily experts in Remote Sensing, nor do they want to be. Instead, they desire to solve problems and make better decisions based upon information products that were derived quickly, intelligently, and in highly automated ways from imagery, frequently involving massive collections of raw data.

Modern GIS systems are facilitating this trend towards simplicity by encapsulating tradecraft, domain expertise, and advanced algorithms within streamlined workflows and Apps that are targeted towards doing just a limited number of otherwise complex functions very quickly and easily, sometimes in near real time, and hence the “Illusion” of simplicity. Put another way, software becomes smarter so the user doesn’t have to. Recent examples illustrating several use cases will be presented, along with observations on near-term trends and continuing areas that need focus and attention.

This presentation will highlight several key research and development activities that are ongoing at ORNL, in particular applications that have been optimized to run in high performance computing environments of any size so they can support the broader geospatial analysis community, along with the challenges and opportunities associated with the optimization process. Four specific activities will be discussed: automated aerial and orbital image registration and enhancement, unmanned aerial systems data collection and processing, the exploration and analysis of spatiotemporal data pulled from a large number of sources with global coverage, and using remotely sensed data to better understand the growth trajectories and internal dynamics of megacities.

At the United States Military Academy (USMA), West Point, New York, students have the opportunity to major in Geospatial Information Science (GIS). For years, the aerial photogrammetry course has been part of the curriculum. Students learn the art and science of photo measurements, photo acquisition, and the mathematical aerotriangulation process that enables highly accurate and precise mensuration. Recently unmanned aerial systems (UAS) have skyrocketed into the public sector for recreational use and commercial exploitation for monitoring, infrastructure inspection, and environmental analysis. As these small, relatively inexpensive aircraft have proliferated, the faculty of the GIS Program at USMA felt it important to maintain the relevancy of the course and add cutting-edge UAS training to the curriculum. This presentation will discuss how USMA has added UAS to the curriculum, what platform we chose, and the results of the inaugural training, which includes the use of ENVI software.

Through a strategic partnership, Harris Corporation and exactEarth have announced a program to bring real-time ship tracking from space to the market in 2018. This presentation will examine the big data management and processing implications of such a capability and will examine the potential applications this will enable across a number of market segments. The presentation will then go on to focus on specific analytics that can be applied to such data. In particular the way in which such data and information services can be fused with optical and SAR capture systems and resulting imagery will be examined to illustrate how this will help provide real insight into behaviors and movements globally across the maritime domain.

Agricultural communities face a number of biological, environmental, and regulatory challenges on a regular basis, ranging from governmental pressure to reduce chemical inputs, to adequate implementation of pest management strategies. Unmanned aerial imagery coupled with large microscale datasets that incorporate a wide array of ground-truthed biological and environmental variables, have the potential to provide improved decision making tools and ease of management to growers. Multiple case studies are presented to examine potentially feasible applications of aerial imagery as it pertains to pest and production management, while existing and emergent technologies are examined as a means to satisfy best management practices effectively reducing the regulatory burden on growers.

Geospatial analytics provides unique capabilities for detection, identification, and characterization of entities from remote sensed observations. Cloud environments and Big Data capabilities are radically changing the classical domains of geospatial analytics and data fusion. Open standards are key to analytics in loosely-coupled ecosystems.

Processing of remote sensed data along with fusion with other spatial information creates new geospatial entities such as signatures and links to other spatial information. Analytics on spatial, temporal and contextual characteristics leads to discovery of new patterns in data. Standards are needed for the accurate and efficient handling of information needed for analytics.

Frictionless information exchange in the context of loosely-coupled and high-cohesion systems requires a different focus when considering big data and cloud services environments. Interactions between cloud-hosted applications requires a focus on data formats and semantics, e.g., for describing algorithms. Interfaces in a big data/cloud environment, e.g., APIs and services for users, become facades on ever-greater capabilities. While at the same time the use of containers allows developers to consider micro-services that differ from previous service oriented architecture (SOA) approaches.

This talk will include activities of the Open Geospatial Consortium (OGC) that are advancing geospatial analytics in cloud and big data environments.

Harris Corporation recently began a new Water Campaign aiming to bring Harris' long history of solving customer's hard problems to bear on a growing societal problem. This presentation will outline recent developments in the Water Campaign and how we are using ENVI, Helios, and other Harris tools to develop cutting edge solutions for modernizing National Meteorological and Hydrometeorological Weather Services, identifying and predicting poor water quality, and developing solutions for enhanced food security.

Planet is building the world’s largest constellation of earth imaging small satellites. We will introduce the development of the project to date and our road map for completing the constellation and making this vast amount of data readily available for analysis at scale. Already, the Planet constellation is producing a vast amount of data from space, and we will discuss how we have had to re-imagine and rebuild geospatial processing and storage to cope with this continuous live stream.

The International Space Station (ISS) is a testament to human ingenuity and innovation, serving the international community as a truly unique facility for conducting space science and an invaluable proving ground for space exploration. The ISS also provides exceptional capabilities and benefits as a customizable multi-modal platform for Earth observation. Advantages of the ISS for remote sensing include the spacecraft infrastructure itself, which delivers power, communications, onboard storage and data downlink, the availability of multiple internal and external sensor mounting points, and even opportunities for crew interaction and repair. Given these resources, the ISS provides an ideal platform for remote sensing technology demonstration as well as a robust framework for co-mounting multiple sensors with different modalities and configurations. As an example, consider the application potential and data fusion opportunities that could be realized through the simultaneous acquisition of hyperspectral, LiDAR and SAR imagery from the same orbiting satellite platform. We present an overview of ISS capabilities for Earth observation, summarize current and planned sensor systems, and discuss how this integrated remote sensing ecosystem can lead to new geospatial insights and analytic capabilities.

Discover how Icaros and Harris Geospatial Solutions partnered to combine their technologies, OneButton™ and ENVI® to provide a complete and continuous workflow for processing and analyzing image data collected from either Unmanned Aerial Systems (UAS) or manned collection platforms. Learn how images are easily and quickly transformed into georeferenced 2D & 3D maps and models that can be processed by OneButton™ for analysis in ENVI®.

Discover how this new solution provides new, highly-detailed, application specific analytics to enable customers to solve some of their most challenging problems ranging from agriculture, natural resources to environmental and urban management. Working with OneButton™ discover how the ENVI® OneButton™ platform will host a new generation of applications for the worldwide UAS market. Learn how this solution, coupled with the evolution of inexpensive miniaturized sensors, will make advanced analytics from aerial imagery easier and cheaper to acquire than ever before. Discover how ENVI® OneButton™ outputs such as Orthomosaics, Digital Elevation Models, true color 3D point clouds, orthorectified multispectral mosaics, and 3D photomesh can be analyzed in ENVI.

In the last few decades, field spectrometers have gone from being heavy and slow, taking a few minutes to collect a single spectrum, to being much more portable and fast. Tens of spectra can now be collected and averaged in a few seconds.

Collections could start with gathering reflectance spectra of materials of interest to first see if there is enough spectral diversity to allow mapping various materials. Many spectra may be needed to make sure that the variability of materials of interest is adequately encompassed. A spectral library may help with the research. Field spectra may also be resampled to match the band passes of overflight data.

For hyperspectral surveys, reflectance spectra of ground calibration targets can be vital to the success and accuracy of a program. Field measurements can be used to convert airborne or satellite data to reflectance and also to remove artifacts.

Finally, if there are any objects that cannot be identified in the course of overflight data analysis, ground reflectance spectra of these objects can help with their interpretation and identification.

Thanks to the connections made at this conference last year, the Algorithm Marketplace is filling with 3rd Party insights for DataMapper users. This presentation will highlight what problems we are trying to solve and how we are solving them using an open IP approach. While we do make use of other analysis engines, we are expanding our relationship with Harris to make better use of the Geospatial Services Framework across our brands.

What if you could use a time machine for the planet and combine it with near real-time information? This is not a dream, it is where the industry is heading. The confluence of increased availability of imagery, lightning fast cloud computing, and advancements in deep learning is making this a reality. The same deep learning algorithms that are used in self driving cars can now be applied at scale against petabytes of data. Although machines aren’t perfect, this is the side of history we need to be on as investments in artificial intelligence pay off.

We have partnered with Harris/ENVI to provide the broadest range of analytic tools available to the community, all on one cloud platform. From vector extraction to atmospheric compensation, or monitoring security perimeters to tracking assets, we have made 100 petabytes of data available for exploitation on our GBDX platform. Putting the compute next to the data allows us to find patterns, identify change, classify materials, identify objects, and inventory structures from a ridiculous amount of data all from the cloud. We can’t do it alone, this is going to take a community.

The new geospatial economy is emerging as a diverse ecosystem of data providers, application developers, integrated computing resources, and innovative new services and products. At the same time, the business models and analytic approaches that drive this economy are also expanding and redefining the market itself. With the rise of this new economy comes exciting opportunity for collaborations on delivering advanced geospatial products to a world-wide community. Here we present a case study of how CloudEO and HySpeed Computing have partnered in one such collaboration. We utilize the CloudEO geo-sharing platform, and its associated computing infrastructure, to implement applications for flood inundation and food security developed by HySpeed Computing to provide global services and analytics derived from Landsat-8 and Sentinel-2 imagery. To achieve these objectives, we leverage the inherent geospatial capabilities and scalability of ENVI and the ENVI Services Engine as the foundation for managing our image processing workload. We highlight the main design steps involved in developing this system, discuss the experiences of integrating applications on the CloudEO geo-sharing platform, and review the benefits such collaboration provides to the customer.

Geospatial data will increasingly be an essential part of solving problems and driving efficiencies into markets and consumers. Success in solving these next generation problems depends on bringing all relevant data sources, processing platforms and enabling client technologies into one aggregated platform. Platforms must be able to reach beyond their own inherit capabilities to have a chance at success in this rapidly changing, yet rewarding marketplace of ideas and implementation. This presentation will detail the building out of the Harris Geospatial Platform that is being built to evolve a legacy of data source aggregation into a cloud-based platform. Emphasis will be placed on providing the essential building blocks of the platform based on best-in-breed data sources and data processing platforms. These building blocks are often required to reach out to other vendor platforms for access and processing to bring everything together through one single interface.

On 19 March 2015, Amazon Web Services (AWS) announced Landsat on AWS, an initiative to make data from the U.S. Geological Survey’s Landsat satellite program freely available in the cloud. Because of Landsat’s global purview and long history, it has become a reference point for all Earth observation work and is considered the gold standard of natural resource satellite imagery.

Within the first year of Landsat on AWS, it delivered over 1 billion requests for Landsat imagery and metadata, globally. Availability of the data in the cloud has led to new product development by companies and startups including Mapbox, Esri, CartoDB, MathWorks, Development Seed, Trimble, Astro Digital, and Blue Raster.

The model of staging data for analysis in the cloud established by Landsat on AWS has since been applied to high resolution radar data, European Space Agency satellite imagery, and EPA atmospheric pollution models.

This session will provide an overview of lessons learned throughout these projects. It will demonstrate how cloud-based object storage is democratizing access to massive publicly-funded data sets that have previously only been available to people with access to large amounts of storage, bandwidth, and computing power.

The ability to acquire new imagery quickly and affordably is a must for customers who want to explore imagery for methods to lower overall farming costs and improve crop yield. Multispectral imagery from last year provides historic context, but does little to help in the current growing season. Growers need to quickly task and affordably acquire data over their fields of interest at regular intervals to make their investment worth the cost. Airbus DS’ One Tasking and customized tasking services provide an affordable option. The Pléiades and SPOT sensors are spatially and spectrally well calibrated so users can acquire imagery over small areas (hectares/individual farms) or large areas. In this presentation, Airbus DS will cover the advantages of frequent collections and explain how daily acquisitions across multiple areas of interest are possible. Harris will show what is possible with a time series of data over one area with its ENVI Precision Ag Tool Kit for extracting crop counts, and individual plant health over time from Pléiades pan-sharpened data.

Unmanned Aerial Systems (UAS) are beginning to transform the electric and gas utility industry. Environmental Consultants Inc. (ECI) and FLōT Systems partnered to explore whether long-range aircraft, capable of beyond line-of-sight operations, combined with remote sensing technology and analytics can assess critical infrastructure more safely and efficiently than traditional inspection methods. Recently, ECI and FLōT Systems successfully executed the first Beyond Visual Line-of-Sight (BVLOS) UAS mission on electric transmission facilities near Amarillo, Texas with authorization by the FAA. We will update the attendees on the current status of initiatives and the business case for commercial UAS applications within the current FAA regulatory framework.

EdgeData equips enterprises to profit from big data intelligence and analytics with a focus in the energy and Unmanned Aerial Systems (UAS) industry. EdgeData is succeeding on its path to extract value from aerial inspections with its repeatable process of Capture – Compute – Consume SM. EdgeData is advancing the wind industry with BladeEdge©, a software portal that transforms raw data from turbine inspections into a dashboard of actionable intelligence for life cycle management of wind energy assets.

Optical Remote Sensing in the Amazon is very much limited to the dry season (May through September), which is why the Brazilian government is implementing a deforestation detection system in the Amazon with the use of orbital radar images. The monitoring of 950,000 square kilometers of the Legal Amazon Region demand analytical capabilities and automation strategies which make ENVI+IDL+SARScape a perfect fit for this task.

Fusion strategies have been widely adopted in data processing to achieve better performance, and we are now using spatial and spectral information fusion to apply these concepts to improve classification of hyperspectral imagery. Our approach treats the spatial and spectral information, collected through feature extraction methods, as nodes in a feature space and utilizes a graph-based data representation to fuse the information and produce the classification. Here, we will talk about spatial and spectral information fusion from hyperspectral imagery, but future work will incorporate spatial information from other modalities such as LiDAR as well.

Coastal southcentral Alaska's Bering and Malaspina Glaciers are the two largest temperate glaciers on Earth. Being sensitive temperate glaciers, they both respond to their dynamically changing environments by thickening, thinning, advancing, and/or retreating on time scales ranging from daily, to annual, to decadal, to centennial, to millennial. A critical question is how both are responding to post-Little Ice Age climate which may in itself be influenced by anthropogenic factors. To address this issue, a longstanding USGS investigation is fusing space-based, airborne, marine, and terrestrial geospatial imagery and data from more than two dozen 'traditional' and 'non-traditional' sources to address the complex relationships and dynamics between these glaciers, their physical environments, and external driving forces such as changing weather and climate. The geospatial data resources of both glaciers are extensive. Parts of each were mapped in the late 19th century and photographed as early as the beginning of the 20th century. Post-World War II, both were imaged and topographically mapped with aerial photography. Each has been completely imaged for 44 years by Landsat. The resulting data analyses document complex glacial and climate histories and an unusual relationship between the glaciers' surface and bed morphology and their remote sensing characteristics.

The recent years are characterized by an increasing deluge of SAR data acquired in a regular and extensive manner over the whole Earth. But data are not yet actionable information, and tools are needed to steer the ship of analytics through these turbulent waters, following the evolution of the SAR measures along time, improving the data interpretation through appropriate processing and filtering, and finally converting the complex backscatter values into geo- and bio-physical parameters and thematic maps.

This presentation will discuss operational cases, carried out exploiting SARscape, where geo-information has been derived from SAR data to have direct use in different application domains, including agriculture, environmental and infrastructure monitoring.

A workflow analytics platform, AppSymphony AppSymphony is a self-service Big Data platform for research analysts and data scientists to easily build applications that can run in private and public clouds. It automates the process of accessing and formatting new data sources, gathering and mixing social media and sensor data, trying and testing new algorithms and remote orchestration across multiple clouds. Based on an intuitive graphical interface, AppSymphony allows end-users to mash-up and integrate data sources in rapid and powerful ways that were previously only possible through the resource-intensive use of data scientists or software engineers.

Presentation abstract: Deep Learning (DL) is a branch of machine learning that has emerged in the past decade and demonstrated unprecedented performance in a range of pattern discovery and recognition tasks in vision, speech and natural language understanding. For example, in 2015 DL outperformed humans in both the recognition of 1000 classes of objects in open-source internet imagery and the transcription of Mandarin Chinese speech to text. DL methods and technology are readily applicable to the problem of learning and then recognizing geospatial signatures from remote sensing data and they address many of the issues that reduce the performance of traditional pattern recognition methods, such as natural target variation, data noise and weak signatures. Furthermore, DL naturally scales across problem types and growing data volumes. We will demonstrate early success in the application of DL to object recognition in wide area remote sensing imagery and full motion video. We will describe how DL methods will naturally extend to other remote sensing modalities such as spectral data. Finally, we will show how through GPU acceleration DL can process massive volumes of incoming sensor data in real-time to recognize signatures and patterns that can populate geospatial databases for downstream analytics.