The "CORONA Digital Atlas of Greece" is an interactive database of the earliest space-extracted photographs of Greece. The Atlas incorporates CORONA satellite images, a reconnaissance program of the US government that was in operation from 1959-1972.
After being declassified by the US government in 1995, the value of CORONA satellite imagery for landscape analysis was quickly recognized by scientists in various disciplines. Although developed for military purposes, the serendipity effect of the enormous dataset created exceptional opportunities for studying the environmental and anthropogenic characteristics of the Earth’s surface, as it was in the middle of the twentieth century, at a reasonably high spatial resolution. In Greece, for example, landscape modifications from urban growth, industrialization, climate change, rural land reclamation, and invasive farming practices, just to name a few, are oftentimes so extreme that surface features from earlier decades are distorted or largely unrecognizable today, if not permanently destroyed. As such, the Atlas is an important and unique instrument for research purposes of the characterization of post-war Greece and the effects of modernization.
From the archaeological point of view, the CORONA Digital Atlas of Greece will help researchers to re-evaluate the status of known archaeological sites and will offer means to study spatial organizations of settlements, especially for the ones which were greatly altered after 1972 or totally destroyed, and thus, were not subject to a scientific study. The wide coverage of the Atlas will also allow new discoveries of previously unknown sites.
CORONA was developed as part of US intelligence efforts during the Cold War to gain large and repeated land coverage of sensitive regions and to supersede the high altitude, yet dangerous U-2 aerial reconnaissance missions. The American satellite program competed directly with Soviet military satellites (Zenit) in what would become a period of intense spaceflight competition between the two world powers. Some CORONA missions (KH4s) employed satellite systems that used two panoramic camera lenses, a Forward and Aft, tilted at 15°. Such an array produced pairs of overlapping photographs on elongated filmstrips that captured about 17-20 km x >230 km of the Earth’s ground surface. The dual camera setup also enabled stereo viewing.
The Atlas is a compendium of 193 individual CORONA frames covering approximately 381,847 km2 of the southern Balkan Peninsula. This total includes land and sea regions of Greece but also parts of Albania, Bulgaria, North Macedonia, and Turkey. The total coverage of the Atlas is nearly doubled (758,012 km2) when combining the cumulative sum of overlapping frames, which may include Forward and Aft images from the same mission or images from entirely different missions and extraction dates. Within Greece, the Atlas provides 91% coverage of the country’s total land mass (120,316 km2 out of 132,091 km2). The only regions of the Atlas that lack coverage are Crete, the southern Dodecanese Islands including Karpathos and Rhodes, and southern Euboea. In general, these areas had limited or no coverage from the CORONA missions, and, in particular, the better resolution KH-4A and KH-4B satellite systems.
The United States Geological Service (USGS) is the repository for CORONA satellite images. Using the USGS Earth Explorer, the Atlas’ collection was acquired at a cost of $30.00 per frame. Imagery is delivered in TIFF digital format after being scanned at 7 micron (3600 dpi) using a photogrammetric scanner. Since the 70 mm format of the photographic negatives are elongated (70 mm), each image filmstrip is scanned and delivered in four parts.
The CORONA imagery that comprise the Atlas range in dates from 1961-1972, having been extracted during different missions over the course of the CORONA program. The majority of images, approximately two-thirds, are from 1965 and afterwards. The at nadir resolution quality is best with the KH-4A (2.74m) and KH-4B (1.83m) satellite systems in use after August 1963. Approximately 80% of the Atlas uses imagery from these two CORONA satellite systems. At the time of CORONA’s declassification in the mid-1990s, the best commercial Earth observation satellites, such as Landsat and SPOT, still had inferior resolution qualities compared to CORONA. It was not until IKONOS (1999, 1.00 m panchromatic resolution) and Quickbird (2001, 0.61 m panchromatic resolution) that publicly available Earth observation satellites began to improve upon the spatial resolution of CORONA, an incredible circumstance considering that the CORONA program predates IKONOS and Quickbird by circa 30 years or more depending on the mission.
The CORONA Footprints layer contains the areas covered by CORONA Images in the CORONA Digital Atlas of Greece. At lower scales, only mission footprints are shown to depict separate CORONA missions for easier viewing.
The Corona Images layer contains all Corona images available for viewing and download. For efficiency this layer is only enabled below a certain zoom level, and thus is not available when the CORONA Digital Atlas of Greece is initialized.
The transparency slider allow the user to adjust transparency of a CORONA imagery. When the slider is on the right side, CORONA imagery is visualized with no transparency. As the slider moves to the left, the CORONA imagery becomes transparent and the modern satellite imagery becomes more visible.
The swipe slider divides the area into two viewing parts, the left side view, dedicated to CORONA imagery, and the right side view, for the modern imagery. The swipe slider allows the user to view the underlying modern satellite imagery without actually turning CORONA image layers on and off, by dragging the swipe slider handle.
The user is also able to draw and measure location, distance and areas on top of the CORONA Digital Atlas of Greece. Location measurements are reported in degrees (World Geodetic System 1984), while distance and area measurements are reported in meters and square meters, respectively.
Distance measurement
Area measurement
Location measurement
The elongated filmstrips present certain challenges in the integration of CORONA satellite images into a digital atlas. For one, there are significant surface distortions in each strip, particularly along the edges, because of the vast spatial coverage and tilt of the panoramic camera. This renders mountainous regions but also valleys near the periphery at exaggerated perspectives that must be corrected. Another issue is the difficulty in achieving correct ground registration (georectification) of the product. A CORONA satellite image does not have spatial or coordinate information for immediate integration into a digital atlas; instead, it must be georeferenced with ground control points (GCPs). However, the CORONA filmstrips are quite large, which makes conventional geographical transformations using polynomial corrections or spline algorithms (common in GIS and imaging software) a particularly laborious task, which is also not that accurate as the photogrammetric solutions. Lastly, the sheer number of CORONA images required to create a near-complete atlas of Greece necessitated a solution that would streamline the data processing.
In order to mitigate these challenges, the Corona Atlas and Referencing System Sunspot of the Center of Advanced Spatial Technologies (CAST), University of Arkansas was utilized at length. The CAST team has developed an open-access online platform for the processing of CORONA satellite imagery using photogrammetric algorithms calibrated to the specific characteristics of CORONA. Although the platform was specifically developed for the KH-4B satellite missions, it works just as well with the earlier missions. The first step is to upload the four raw satellite datasets that comprise a single CORONA scene onto the Sunspot platform, which then reconstructs a complete scene using an automated scale-invariant feature transform (SIFT) algorithm. The second step is to establish fiducial points at the corners of the scene. As a third and final step, the user is required to match approximately 15-20 ground control points (GCPs) by using a custom digitizer tool that displays side-by-side frames of the CORONA scene and todays satellite and aerial imagery (e.g. ESRI World Imagery, Greek Land Registry). Sunspot enables the exporting of the final product in National Imagery Transmission Format (NITF) format along with an SRTM Digital Elevation Model (DEM) of the area of interest. With the DEM, it is possible to further orthorectify the CORONA scene in GIS or other imaging software. While these were the normal steps for processing the individual CORONA scenes, some issues arose that required the implementation of alternative processing methods. Since the majority of satellite images of Greece often include extensive sea areas with no surface distinctions, at times Sunspot was unable to generate reliable GCPs and complete a scene using the automated SIFT algorithm. In other instances, cloud coverage or other land features produced alignment errors. In such cases, Sunspot could no longer be used and polynomial corrections in GIS were applied for ground registration.
The Atlas of Greece was inspired by the CORONA Atlas & Referencing System for the Middle East region. The Center for Advanced Spatial Technologies (CAST) at the University of Arkansas kindly provided access to the main CORONA Georeferencing System (sunspot) for the proposed project. Their personnel provided valuable technical expertise for the use of their tool.
This work was supported by the project "POLITEIA II" (Politismos-Technologia, New Technologies in the Research, Study, Documentation and Access to the Information for Cultural Heritage Objects and Monuments II) (MIS 5002478), implemented under the "Action for the Strategic Development on the Research and Technological Sector" and funded by the Operational Programme "Competitiveness, Entrepreneurship and Innovation" (NSRF 2014-2020) and co-financed by Greece and the European Union (European Regional Development Fund).