-Using the DEM from the 2000 and 1996 AIRSAR data, we generated a difference image (after removing some interference effects) which appears to show large patches where the tree canopy height had significantly reduced. This is of particular interest because a large cyclone had caused damage to the Wet Tropics area between acquisition dates. We now need to confirm this using aerial photography.

-During the PACRIM field campaign, the CSIRO/NASA/Uni of Queensland team placed a number of tone generators under mangrove and rainforest canopies in an attempt to quantify the amount of radar penetration into these dense vegetation types, for different radar wavelengths. Some were also placed in an open area to provide a calibration response. After finding their locations in the imagery, it appears that the tone generator responses, apart from the calibration, are not visible in the processed data. Whilst this was somewhat disappointing, it is also an indication of the density of mangroves and rainforest. Due to time limitations, Ernesto is now going to look at the tone generator responses in the raw radar data. We then hope to relate these values to our field measurements of the vegetation.

I also had the opportunity to meet with Sasan Saatchi at JPL who is working on the GRFM (Global RainForest Mapping) project. I was able to learn more about their work on mapping large regions of rainforest using JERS radar data, which is also what we are doing in Far North Queensland, and discuss their mangrove work.

Whilst I was also hoping to look at some of the SRTM (Shuttle Radar Topographic Mission) data and discuss its availability, the September 11th terrorist attacks had changed its processing priorities and lead to uncertainties about the date and format of its release to the public.


Third International Conference on Geospatial Information in Agriculture and Forestry, Denver Colorado
5-7th November 2001

This international conference focuses on advances in geospatial information technologies in relation to agricultural and forestry information needs. Participants were from agricultural and forestry businesses along with technology companies and research institutions. At this conference I presented a paper titled: “Mapping Tree Crowns using hyperspectral and high spatial resolution imagery” (Authors: Catherine Ticehurst, Leo Lymburner, Alex Held, Claudia Palylyk, Darrell Martindale, Stuart Phinn and Michael Stanford). It included our algorithms developed for delineating and characterizing complex vegetation environments, such as rainforest canopies.

There were around 150-200 attendees. Topics included GPS, GIS, LIDAR, Hyperspectral, Radar, and forestry assessment. Due to the increasing interest in LIDAR technology for forest measurements, there was a special session on this topic. It included presentations on interesting LIDAR profiles for looking at subtle changes in forest structure, isolating individual crowns of a leaf-off deciduous forest, and on generating a virtual forest for the user walk around in (virtual compass included). The USDA Forest Service also has a useful forest Stand Visualisation Software (SVS), which can be downloaded from http://forsys.cfr.washington.edu.

There were also a number of application based presentations in the area of precision agriculture, where ground based sensors were attached to the front of tractors, and the fertilizer/weed spray was accurately applied out the back in real time. The use of hyperspectral sensors (such as CASI) was presented in a number of papers for detecting weeds, stressed vegetation and target species. This included an interesting paper by Doug Davison (et al) from ITRES who were using CASI (36 bands) for early detection of pine trees affected by the mountain pine beetle. They did this using an automated crown delineation and detection algorithm, which appeared to be able to separate the healthy pines from the infected pines still in their green-stage.

Conference abstracts can be found at: www.erim-int.com/CONF/conf.html