A Burmese python, ready for its close-up. Recent work from U.S. scientists suggests that near-infrared cameras could help in the effort to find and neutralize the snakes, an invasive species that has emerged as an ecological threat in the Florida Everglades. [Image: Hillary Kladke / Getty Images]
Since the late 1990s, invasive Burmese pythons—giant snakes that can weigh in at more than 90 kg (200 lb)—have emerged as a potent environmental threat in the Florida Everglades, depleting natural species there. But the well-camouflaged snakes have been hard to ferret out and neutralize amid the natural foliage. Now, a research team from the University of Central Florida College of Optics and Photonics (CREOL), USA, thinks it has found the Achilles’ heel of the reptiles: They show up much better in near-infrared light (Appl. Optics, doi: 10.1364/AO.419320).
(Enormous) snakes in the grass
Burmese pythons, born as 50-cm (20-inch) hatchlings, grow to a length of 2 m (7 ft) or more in a year and can reach lengths as great at 7.9 m (26 ft) in the course of their lives. They first started showing up in the United States in the 1980s, imported from Asia as exotic pets. Then, in 1992, Hurricane Andrew destroyed a python breeding facility, accidentally releasing the snakes into the Everglades.
The habitat proved a felicitous one for the pythons, which have bred and proliferated since then and are now widespread not only in Everglades National Park (ENP) but in much of southern Florida. And they appear to be on the move, expanding northward in Florida. Should the huge snakes gain a stronghold there, scientists think that they could eventually spread to other zones with suitable climates—perhaps as far distant as Texas to the west and Virginia to the north.
The pythons’ natural camouflage makes them very difficult to pick out amid the grass and foliage of the Everglades. But they’ve definitely made their presence known in other ways. A 2012 study found that in the southernmost areas of ENP, where the invasive snakes have been established the longest, populations of raccoons and opossums had crashed by 99% since 1997, and species such as marsh rabbits, cottontail rabbits and foxes had essentially vanished. And the same animals were routinely being found in the stomach contents of captured Burmese pythons from the park.
These ecological impacts have made rooting out and neutralizing the pythons a top priority. The state of Florida has even encouraged human hunters to go after the snakes—with the 2021 hunting season beginning just this month.
But finding the pythons, which blend so effectively into the Everglades foliage, turns out to be a devilishly difficult proposition in visible light. The team behind the new work—led by assistant professor C. Kyle Renshaw and Ph.D. candidate Jennifer Hewitt at CREOL, and also including Orges Furxhi of Imec USA and OSA Fellow Ronald Driggers of the University of Arizona, USA—wanted to see if the snakes might show up better in another part of the spectrum.
NIR versus visible
Work published in 2019 by Driggers, Furxhi and others, employing hyperspectral cameras in the lab, had already fleshed out the reflectivity of pythons relative to natural Florida foliage in a number of spectral bands, and revealed that the contrast between the snakes and the natural background increased at wavelengths longer than 750 nm. In the new work, the researchers used that finding as a launchpad for a test to see if near-infrared (NIR) detection of the beasts could work in a natural setting.
The team’s setup included visible-light and NIR CMOS cameras (center) for both daytime and nighttime imaging, and halogen lamps (top) and NIR LEDs (bottom) for nighttime illumination. [Image: J. Hewitt et al., Appl. Opt. 60, 5066 (2021)]
To find out, the team designed and constructed a CMOS camera, with a visible–NIR, fixed-focal-length lens, that sensed at wavelengths around 850 nm. Hewitt then went into the field and snapped pictures of pythons in different locations with different background scenery, using the NIR camera and another, visible-light camera that differed from the first only in its spectral coverage. She collected the pictures both in the daytime, using natural sunlight, and at night, using both visible (halogen) and NIR (LED) artificial illumination.
The researchers then had individual volunteers examine the photos to determine whether they could spot pythons in the images, and to click on suspected pythons in randomly presented images (some of which did not actually include any snakes). The results suggested that the extra dose of contrast in near-infrared light was enough to allow participants to detect pythons at distances 20% farther away than with the use of visible light alone. What’s more, using the NIR bands provided an edge for detection both during the day and at night.
Toward automated detection?
The team is now working with the Florida Fish and Wildlife Conservation Commission (FWC) to build on these results, and test more rigorously whether the approach might work in practice. Team leader and coauthor Kyle Renshaw suggests that the NIR cameras “could be mounted on trucks or drones” to help find the elusive beasts. And coupling the digital-camera data with computer algorithms, he says, might even open up the prospect of an automated detection system. If that idea pans out, such a sytem could be a potent tool in the quest to stop the invasive snakes in their tracks.