Advanced airborne technologies for mapping and monitoring native Australian vegetation

Citation

Land & Water Australia. 2009. Advanced airborne technologies for mapping and monitoring native Australian vegetation. [Online] (Updated July 3rd, 2009)
Available at: http://lwa.gov.au/node/2966 [Accessed Tuesday 26th of April 2011 03:43:08 AM ].

Airborne sensors can be used by natural resource managers and researchers to collect visible (vegetation cover) and invisible (such as canopy condition or water use) data.

Data can then be used to derive estimates of carbon storage in vegetation, fire fuel load, biodiversity, plant growth rates, occurrence of invasive species, and many other measures. This proof-of-concept project used full waveform resolving airborne Lidar combined with other airborne technologies (multi- and hyper-spectral scanners, digital video and stills) flown simultaneously on a small and cost-efficient aircraft. The project focused on instrumentation, calibration and airborne data acquisition and integration. Numerous field trials and collaborations Australia-wide demonstrated that the technology is of high value to scientific applications in assessing and monitoring the natural environment.

Project Objectives

To resolve the hardware issues of integrating the two imaging sensors on a single aircraft. . To develop the software tools for combining hyperspectral and ALS data. .
To validate the proposed technology in a range of different environments .
To demonstrate the new technology to the natural resource management community in Australia.

Project collaborations and data gathering

Of 25 target areas/sites flown in 2007, data has been analysed and evaluated further for the following four projects:

1. Performance of vines under irrigation (in collaboration with B. Loveys, CSIRO Plant Industry)
The AISA+ hyper-spectral imagery was compared to in-situ measurements of stomatal conductance of grape vines under different irrigation treatments in the South Australian Riverland at Waikerie. Hyper-spectral image spectra revealed differences in canopy reflectance between the irrigation treatments and supported the in-situ measurements of canopy performance.

2. Discrimination of variation in composition of native and regenerating vegetation (in collaboration with J. Allen, D. Paton and B. Ostendorf, University of Adelaide)
Research is being conducted by an Honours student (July 2008 – June 2009) to evaluate the combined use of airborne hyper-spectral imagery with Lidar to characterise native vegetation composition and structure. The project is examining differentiation of hyper-spectral signatures amongst species, use of the imagery for mapping species, and development of approaches to the integration of vegetation information derived from Lidar and hyper-spectral data. The study is based in remnant and regenerating vegetation at Monarto, South Australia. This project is being conducted in parallel with PhD research (J. Allen) which is quantifying habitat diversity and value in regenerating vegetation; it is hoped that image-based vegetation mapping will provide new methods for characterising vegetation diversity.

3. A physical model-based approach for interpreting hyper-spectral imagery of eucalyptus woodland in South Australia (in collaboration with M. Mõttus and M. Rautiainen, Tartuu Observatory, Estonia)
This project is testing the feasibility of a physically-based approach to determine the structural properties of natural eucalyptus canopies at the Monarto test site in South Australia. The airborne AISA NVIR hyper-spectral image acquired during the summer of 2007 was processed to top-of-canopy reflectance. Ground truth measurements were carried out in October 2008 to determine basic canopy parameters (tree density, canopy height). Cover photography, measurements of crown cover, crown porosity and leaf area index of the 1.5 ha test area were used. Canopy reflectance was parameterised using photon recollision probability theory.

4. Mapping and monitoring vegetation associated with mound springs, Great Artesian Basin (National Water Initiative project 2008–2010, M. Lewis and D. White)
Hyper-spectral, thermal and Lidar data collected during the 2007 field campaign are being evaluated as tools for mapping the composition and structure of perennial wetland vegetation associated with mound springs in northern arid South Australia. Spectral signatures to characterise the vegetation and to assist with image analysis were collected in February 2008, with full analysis and evaluation of the airborne data being conducted by postdoctoral researcher D. White during 2009. Full evaluation of the data from some of the sites is still on-going and will continue into the future. However, there is now sufficient validation data available to confirm that the technology demonstrated is of high value to scientific applications in assessing and monitoring the natural environment.

The technology (or part thereof) introduced and demonstrated in this project has been applied in several other research projects. Amongst them are (in most cases, the hyper-spectral scanner was not available, so the ARA-owned tri-spectral scanner was used as a substitute):

ARC-funded large study of the northern Australian savanna (Lidar, hyper-spectral and tri-spectral)
Industry-funded study of seabed and reefs in Western Australia (hyper-spectral and tri-spectral)
NRM-funded study of native vegetation in swamps on Fleurieu Peninsula, South Australia (Lidar and tri-spectral)
Land & Water Australia-funded study of vegetation and gully erosion in far north Queensland (Lidar and Tri-spectral)
NRM-funded study of vegetation and land erosion over all of Queensland (Lidar and tri-spectral).