Recent articles outline a general dissatisfaction with geospatial innovation in the National Geospatial-Intelligence Agency. Part of the issue, contracting for commercial imagery purchases, resides with the National Reconnaissance Office. Both agencies face challenges in dealing with the volume of geospatial data from space, increasing mission complexity and incorporating software, hardware, and data being created by the smallsat industry.

How both agencies determine price and value for the geospatial dollar and the geospatial data remains unclear to industry, oversight and other government agencies. Once, in an earlier time of big data, both agencies cooperated innovatively in creating a scale that removed subjectivity and brought clarity to an important aspect of satellite imagery.

In 1971, a new satellite system, the KH-9, was launched. Its scanning camera returned 16-times the amount of data on film than the KH-4, but parts of each KH-9 scanned frame had varying utility for the imagery analysts. Along with the vastly increased and variable data, a new national intelligence mission arose at this time, treaty monitoring, based on the ongoing SALT negotiations. This new mission meant that negotiators and monitors outside the imagery intelligence profession had to understand resolution, or what could be expected to be seen on an image. And, the National Reconnaissance Office (NRO) was beginning to develop a revolutionary new kind of satellite, and imagery analysts and others would soon have to compare chemical photography with digital imagery. All these developments required a new kind of measurement.

Two leaders, John Hicks at the National Photographic Interpretation Center (NPIC) and Bob Kohler at the NRO, created a joint team in 1972 to address these data, mission and future innovation challenges. The team created a distinction — “Interpretability,” and a scale NIIRS (The National Imagery Interpretation Rating Scale) that used objective correlatives to define differences in interpretability.

A Hexagon KH-9 reconnaissance satellite being moved into a building at the National Museum of the U.S. Air Force in Dayton, Ohio, in May 2016. Credit: U.S. Air Force photo

To measure the technical resolution of satellite cameras, the photographic interpretation and imagery analysis community had used a distinction, Image quality, measured in ground sampling distance. Image quality refers to the technical achievement of the camera or sensor. It is defined mathematically. From the beginnings of the film return era in 1960, a joint NPIC/NRO team checked the image quality equation as part of the initialization for each new government satellite. The distinction introduced by NIIRS, Interpretability, refers to how useful an image is to helping an imagery analyst answer an intelligence question. It is very possible to have an image of high quality and no interpretability. Hundreds of thousands of cloud covered images meet these criteria.

Until 1972, all photo interpreters and imagery analysts subjectively defined interpretability as poor, fair, good or very good. The NIIRS defined interpretability on a scale from 0 to 9 by using objective correlatives. Each increasing number indicates that more detail on an object can be observed. For instance, on NIIRS 0 imagery, nothing can be discerned; on NIIRS 4 imagery, an analyst could “Identify trucks at a ground forces installation as cargo, flat bed or van; and on NIIRS 7, an analyst could identify a missile transporter/erector (fixed or mobile system) when not in a known missile activity area.” The scale includes humans, as on NIIRS 8 imagery individual limbs can be detected . Correlatives were created for air, navy, ground forces, and missile related equipment and in 1996 a civil NIIRS with cultural or non-military criteria was created.

In 1971, NIIRS provided a way for analysts to measure the different interpretabilities on the new KH-9 satellite scan images; it provided a means for the Strategic Arms Limitation arms control treaty staff to define the interpretability needed to identify a treaty violation, and in January 1977 it provided the engineers, collection managers, and imagery analysts a way to determine the intelligence utility of the innovative KH-11 digital imagery. By 1974, NIIRS certification became part of every NPIC analyst’s training. NPIC sent out NIIRS training teams to certify the DoD imagery schools.

Ten years ago, in 2009, NGA stopped NIIRS certification and then stopped supporting NIIRS training for the military. Image quality is still measured for government satellites and interpretability still matters in geospatial analysis. Other basic concepts useful for government and commercial geospatial collection and analysis—swath width and persistence to name two— also do not have a common units of measure. So if a new vendor claims these capabilities, they will not be measured by the government.

If NGA or the NRO wish to incorporate the new technologies of the commercial smallsat companies, apply innovative ideas to current missions, and advance the quality and trust in image interpretation algorithms, creating common scales of measurement, like NIIRS, across government and commercial geospatial communities would increase geospatial innovation.

Until the intelligence community defines cost and value by measuring geospatial information, it has to rely on one measurement—six inches. The length of an American dollar.


Jack O’Connor is a former CIA and NGA senior executive who spent nearly all his 31-year career in imagery and geospatial analysis. In retirement, he has written for the NRO’s Center for the Study of National Reconnaissance and NGA’s Center for the Study of Geospatial Intelligence. He also consulted with Draper Laboratory on a smallsat study for NGA. Currently, he is the program director for the M.S. in geospatial intelligence at the Krieger School of Johns Hopkins University.

This article originally appeared in the Oct. 7, 2019 issue of SpaceNews magazine.