Earthshots: Satellite Images of Environmental Change

You are here

» Landsat’s Return Beam Vidicon
 

A sensor on the early Landsats had a name worthy of some Star Trek gadget. Originally Landsat’s primary sensor, the Return Beam Vidicon (RBV) flew on the first three Landsats.

A sample RBV image shows the northwestern coast of Madagascar. The black-and-white image from 1981 has higher resolution than the Multispectral Scanner (MSS) on board the early Landsats. So what happened to these RBV images, and are they useful today?

 

What’s a Vidicon?

A vidicon is a television camera tube that formed an image by focusing light onto a photoconductive faceplate. An electron beam scanned the faceplate, detecting light intensity for each scan line. The beam then bounced back by an electrically charged area. The resulting picture was made up of about 5,000 separate scan lines (compared to 525 for a traditional television picture).

Landsat’s RBV had an inauspicious beginning. It rode into orbit on Landsat 1 on July 23, 1972. During orbit 196, just 14 days later, a relay in the Power Switching Module of the spacecraft got stuck in a permanently “on” state.

The problem could have been fixed with a difficult command sequence, but the other sensor on Landsat 1, the MSS, was already showing its excellent performance and became the favored sensor for its capability to acquire near-infrared data. So the RBV on Landsat 1 was not reactivated. In the short time it operated, it recorded 1,692 images.

The RBV camera that flew on board Landsat 3 was redesigned and given a slightly different mission. This RBV had a spatial resolution higher than the MSS, 40 meters, to add a new dimension to the MSS’s multispectral coverage. The higher detail could be used for detailed ground mapping. The Landsat 3 RBV acquired many thousands more images than either one on board Landsats 1 and 2. The RBV imagery is scattered across the globe, and all of it resides in the EROS archive.

Because of RBV’s higher spatial resolution than MSS, glaciologists were able to use the RBV imagery for spotting more detail. “If you want to look at the ice front, you want to get as sharp a picture as you can,” John Dwyer, Chief of the EROS Science and Applications Branch, said. “They could map crevasses and their displacement and movement over time and map the surface velocity. Once you straighten out the geometry of the image, the RBV can still be useful.”

Yes, the geometry of an RBV image had to be “straightened out.” It turns out there are a few problems inherent in the RBV data.

The RBV was more like a television camera than a sensor, but television pictures tended to distort the image. Researchers using satellite imagery want the image to resemble an accurate map. They expect straight lines for latitude and longitude, for example. RBV images, however, were more like maps printed on a rubber sheet. They were easily stretched out of shape, distorting those imaginary latitude and longitude lines. These distortions could be corrected, but that would have increased processing time and cost.

Furthermore, when imaging over Antarctica, RBV tended to go snow blind. Areas of high reflectance, like snow and ice, could be overexposed. Therefore, cloud cover and snow cover were easily confused.

Because RBV’s main purpose was mapping the earth, the imagery included a reseau grid. No other Landsat sensors contain reseau marks. These plus signs on the images are used to correct distortions in the image after development or scanning. As John Faundeen, Archivist at EROS, put it, the reseau marks allowed you to “tie the imagery to the earth to verify the accuracy.”

In all, there were officially 11 inherent problems with RBV data. Other problems include corners out of focus, occasional black vertical lines, and missing or distorted reseaus. These problems are detailed in a 1981 USGS Open-File Report.

The RBV image shown in this section is in Brazil and shows a stairstep anomaly on the right side of the image. This stairstep phenomenon appears in several RBV images.

Roughly 138,000 images were taken by the RBV on Landsat 3. The data were recorded to 70-mm black and white film rolls at the NASA Goddard Space Flight Center and delivered to EROS. This film, stored either on 70-mm film or in envelopes on film “chips,” is part of the EROS archive, and it’s the only copy.

Only 64 of those images have been scanned from their film source for immediate download via EarthExplorer (look under Landsat Legacy in the Data Sets tab). The rest of them are there and you can find their locations, but the film would have to be scanned before you can download a high-resolution version.

If a scene has not been already scanned, users can place an order to scan the film for $30 per scene. Once it’s scanned and a high-resolution digital version made available for download, it’s freely available worldwide. But as noted, any of these images have a decent chance of having geometric distortions.

Despite its problems, RBV data has been used in land change studies.

RBV data was used in a recent scientific study of glaciers in Turkey. Scientists used RBV images from 1980 (along with dozens of other Landsat images and high-resolution commercial satellite images) to document the areal extent of all glaciers in that country. The researchers noted that RBV’s higher resolution was an excellent source for glacier studies.

The glacier covering the top of Mount Ararat is Turkey’s largest glacier. Based on data from the RBV and more recent Landsat sensors, the glacier area of Ararat diminished from 8.9 square kilometers in 1977 to 5.6 square kilometers in 2008, according to the study. Most loss took place on the southern, western, and eastern sides of the mountain.

Even though RBV had an impressive spatial resolution, the near-infrared and shortwave infrared imaging capability on current Landsats defines the extent of ice more clearly than RBV could.

The depth of the Landsat archive across the history of Landsat sensors made this study possible. It’s important to continue to monitor these glaciers and gauge the effects of these changes, and RBV turned out to be an excellent source where data did not already exist.

Mt. Ararat Glacier Extent
YearAREA
19778.9 km2
19878.7 km2
19987.1 km2
20006.7 km2
20046.3 km2
20085.6 km2
20135.3 km2

One of the factors that improves on Landsat sensors over the project’s history is the quality of the data. For example, compare these images of glaciers in Alaska’s Chitina River Valley from 1980. There is more background noise in an RBV image than there is in the MSS image. The RBV on Landsat 3 had a slightly higher spatial resolution, but the low signal-to-noise ratio makes it harder to pull out surface detail.

The 2018 image from Landsat 8’s Operational Land Imager has even better signal-to-noise ratio and higher resolution than both early sensors. The image reveals more detail about the roughness and flow of the ice. Landsat 8’s shortwave infrared bands provide a better distinction between ice and rock or soil.

High-quality data is needed in studying glaciers. Glaciers reflect visible light and appear bright white in natural color satellite images. But a glacier absorbs near-infrared light. The near-infrared and shortwave infrared imaging on later Landsats show the difference between snow and bare ground more clearly.

However, glaciologists still found RBV useful for mapping arctic regions. In high latitudes, images acquired at low sun angles helped enhance local topographic relief. The high resolution of the RBV on Landsat 3 supplemented MSS imagery.

In the RBV image of the Chitina River Valley, the brightest wavy streaks are glaciers. Note that they extend farther downstream in 1980 than they do in 2018. In the Landsat 8 image, wavy blue streaks indicate glaciers. They are generally in the same location as in the 1980 image, just diminished.

While RBV might not be exactly what is needed or have the ideal attributes for a study or for mapping forests, glaciers, and other land covers, perhaps it’s one more tool to add to the collection in studying land change. It’s at least worth another look. Besides, as Faundeen said, “RBV is another interesting chapter in the Landsat story.”

References

Have a question or comment? Please contact us at custserv@usgs.gov