Not just for measuring trees: Laser scanner captures “Billy” the African elephant

Gepubliceerd op
16 december 2013

Thomas Hubert summarizes his talk with Kim Calders at the CIFOR Forests News Blog, as part of the CGIAR Research Program on Forests, Trees and Agroforestry.

Meet Billy the wild African elephant. At 2.4m in height, he is a little smaller than your average African elephant, but he’s become somewhat of a local celebrity to scientists working in Gabon’s Lope National Park.

"He often visits the Park’s research station to gorge on the fruits of surrounding trees", says Kim Calders of the Laboratory of Geo-Information Science and Remote Sensing at Wageningen University who was carrying out fieldwork when he spotted Billy from a distance and set up his laser scanner to get a picture.

Calders and his colleagues Jose Gonzalez de Tanago Menaca also from Wageningen University and Andrew Burt from University College London were in Lope National Park experimenting with a relatively new remote sensing technology called LiDAR.

LiDAR measures distance by illuminating a target with a laser and analyzing the reflected light. The computer-generated image of Billy the elephant is one example of the technology’s versatility; it is already used by police forces to collect evidence at collisions, and by architects and construction engineers on building sites.

It’s also a technology that is increasingly used by foresters to take high-precision measurements of tree diameter, height, volume, vertical structure, canopy cover and branching architecture. Such biomass mapping is critical for estimating carbon storage of forests as part of climate mitigation activities, such as the UN-backed scheme to Reduce Emissions from Deforestation and forest Degradation (REDD+).

Accurate mapping is critical for characterizing forests and linking ground measurements with those from satellites, said Martin Herold, Chair of Remote Sensing at Wageningen University.

"Having a full 3-D reconstruction of the forest is really important for this integration and goes beyond any other approach available so far", he said.  

Forest structure over large areas is best measured using airborne LiDAR says Calders, where laser sensors mounted on aircraft to scan forest areas from the sky.

"Airborne LiDAR is looking to the forest from above the canopy and does not give you so much as terrestrial LiDAR that is looking hemispherically from below the canopy", Calders said.

Virtual Forest

Scientists have been using tripod-mounted laser equipment (known as "terrestrial LiDAR") to record the position of millions of points reflected by the surface of trees in forest plots. Using mathematical models and computer programs, they then join the dots to reconstruct each individual tree.

The resulting "virtual forests"are made up of countless cylinder shapes representing sections of tree trunks and branches, measured with great accuracy.

Calders and his colleagues have compared LiDAR measurements with traditional methods involving cutting down and measuring sample trees. They have found the technology to be extremely accurate: a eucalyptus tree was measured with LiDAR at 0.55m in diameter by 23.9m in height (actual measurements 0.59m by 24m).

"If you have done a good job, the data you get on your computer is really how it would be if you were there in person", Calders said.

Detailed ground data collected on sample plots can also be used to calibrate larger-scale airborne LiDAR measurements.

"By combining them you can get a better picture of overall forest composition that is of interest not only for carbon research but for forest ecologists and other forest scientists", Calders said.

"The interest is growing and so is the forest-related terrestrial LIDAR community".

Improving Computer Models

Scientists are looking for such ways of scaling up the findings of terrestrial surveys, as they remain limited in scope: scanning a 100x100m plot of forest can take up to four days in dense tropical forests, carrying equipment deep into the jungle.

"Improving the accuracy and versatility of computer models to process and analyze large LIDAR datasets is one of the key objectives now", Calders says.

"The validation of these results against independent reference data is a key area that researchers are now working on but the precision and detail that we have seen from such data is unprecedented and hard to measure otherwise", Calders said.