Overcoming the challenges of remote sensing

Ottignies, Ottignies-Louvain-la-Neuve, Belgium. Photo by Antoine Pirard on Unsplash.
Apr 2019
26

A series of live coverage updates from Global Land Programme’s 2019 Open Science Meeting in Bern, Switzerland.

25 April, 10:00 pm CET

Catching images of land use from above can be useful in improving crop yields and tracking deforestation, but it is not as simple as just a few snaps from the “Spy in the Sky”.

Experts in remote sensing at the Global Land Programme’s Open Science Meeting listed complex hurdles to their work ranging from cloud cover to systems that don’t work together to government privacy laws.

It is what session organiser Bronwyn Price called “the nitty gritty stuff in between”.

That said, great strides are being made in the field as the result of advances in high resolution imaging from space, sky and land.

Two projects – one in Portugal, the other in Southeast Asia – show both the uses and the challenges.

The University of Lisbon is looking for ways to help farmers on pastureland target the fertiliser they need rather than simply spread it widely. They are focusing on land that has been resurrected over the past decades through the introduction of natural seeding, turning it from scrubland to richer pasture.

The current goal of the programme is to tell farmers what areas need more phosphorus and what areas don’t. The data is gathered using a combination of satellites, drones and Earth-bound research.

The problem, researcher Ricardo Teixeira says, is that they currently have to use the NVDI (normalised difference vegetation index) to decide where the phosphorus needs to go, essentially judging by the greenness of the land.

What they really want (and are working to achieve) is a data set that allows the aerial “spies” to work it out with greater accuracy.

Teixeira is only half-joking when he says he wants cows to send text messages.

Across the world in the Greater Mekong Subregion, Chinese academics have been trying to overcome issues such as cloud cover and rapidly growing forests to track logging.

The Chinese Academy of Forestry and others have achieved around 86 percent accuracy using image compositions from a number of different satellites, weighting pixels to get closed to what is happening.

But the data collected from different satellites is not always compatible, which limits their investigation. The researchers want to find a way to combine everything.

The forests in the region also change so quickly – recovering somewhat from logging – that the group has taken to capturing images on a quarterly basis to track shifting conditions.

---

Check back for our ongoing coverage of the Global Land Programme’s 2019 Open Science Meeting. You can view the plenary keynotes on the free online livestream from the event, and follow minute-by-minute coverage on Twitter at @GlobalLandP or use the hashtag #GLPOSM to follow attendees' posts.

 

Comments