Proteus FZC, a provider of satellite-derived mapping and classification services, has completed a demonstration project using satellite imagery to inventory tree plantations in the Emirate of Abu Dhabi. The tree mapping pilot is a spin-off of a larger Emirate-wide habitat and land use/land cover (LULC) project now being spearheaded by Proteus.
In the pilot, the Proteus team processed multispectral data collected by DigitalGlobe’s WorldView-2 commercial imaging satellite to identify the species and conditions of individual trees within the pilot area. Proteus managed the project in which GMV of Spain performed image processing and automatic tree extraction with local ground-truthing support from Nautica Environmental Associates in Abu Dhabi.
“This pilot demonstrates the viability of using very high-resolution multispectral data to establish a baseline inventory of tree type and health within diverse forest plantations,” said Proteus Project Manager, Richard Flemmings. “We mapped every tree crown larger than one meter in diameter in the pilot area with minimal ground truthing and delivered the results in an ESRI geodatabase.”
Abu Dhabi has planted forest plantations totalling nearly 20 million trees that provide aesthetic and environmental benefits throughout the Emirate. Comprised of gaff, acacia, mesquite and other species, these plantations require continuous irrigation with desalinated water at considerable expense. The Environment Agency – Abu Dhabi (EAD) requested the pilot as it seeks to find an efficient and cost-effective way to monitor the forest stands.
“These forest plantations are valuable resources for Abu Dhabi,” said Richard. “EAD envisions a monitoring program that identifies isolated health problems so they can be remediated.”
Proteus applied image processing techniques using all eight WorldView-2 multispectral bands plus one panchromatic band, with an emphasis on the mid-infrared spectra. This multi-step processing distinguished several tree species by their individual canopies and detected stress, possibly related to irrigation, salinity or infestation issues, in some of the trees.
Compared to the traditional ways of monitoring forestry plantations in the area, which implies walking along the tree lines to detect damaged/dead trees, remote sensing offers a cost-effective alternative.
“We used pan-sharpened imagery to create the baseline plantation map, but less expensive lower-resolution imagery such as Landsat could be used for ongoing periodic monitoring,” said Richard. “This individual tree mapping technique can be applied to create forest inventories of other species elsewhere in the world.”
Aside from the tree plantation pilot, Proteus is currently engaged in a fine-scale satellite-derived terrestrial and marine LULC and habitat mapping project for the entire Emirate of Abu Dhabi. The three-phase, multi-million dollar project will ultimately include 60,000 sq. km. of land area and the coastal marine environment down to the 15-meter contour. Advanced processing algorithms are being used to extract LULC and habitat features from high-resolution multispectral satellite imagery acquired over both land and sea. Tree plantations are one of the LULC types identified by the Proteus team in the larger EAD project.
In the coastal Arabian Gulf, Proteus is delivering seabed classifications to depths of 15-20 meters. Deliverables for each phase of the Abu Dhabi mapping project include bathymetric analysis, orthorectified mosaic, LULC/habitat ecological classifications, geospatial models, printed maps at multiple scales, and knowledge transfer.
Since 2011, Proteus has been delivering solutions for mapping and classification projects using multispectral satellite imagery. These mapping projects have been delivered for environmental, oil & gas, engineering and other coastal zone applications in Europe, USA, the Middle East and Caribbean.