Remote Sensing

Aerial Photography

Aerial photographs that cover the coastal areas of Tasmania are sourced from the archives of the Department of Primary Industry, Water and Environment. Photographs are selected based on calm water surfaces and suitable sun glint and camera angle conditions for determining seabed features. The photos vary in scale from 1:12 500 to 1: 42 000 and are chosen regardless of lineage or processing as both colour or black and white photographs can be useful. The Appendix section of each report lists the photograph details including the archive number of each photo used. The selected aerial photographs are captured on an A3 colour scanner at 600dpi (dots per inch). The photographs are then stored as 24bit colour images. Each image is georeferenced using ArcInfo (ESRI) to The LIST 1:25 000 (Land Information Systems Tasmania) coastline coverage in AGD66. To rectify, a minimum of 15 ground control points are selected for each image.

Capturing information from Aerial Photographs
The aerial photographs are displayed in ArcView 3.2. In order to clearly identify certain features such as reef and sand, the colour intensity of the image may be altered. For multi-band images such as aerial photographs the user can manipulate the red, blue and green bands to reduce or increase intensity of the colour and contrast of the sub surface habitats.

These aerial photographs are merely used as a secondary source of information to aid in determining the inner boundaries of the habitat type mainly for reef and sand habitats only, and not for primary mapping of habitat boundaries.

The use of pre-existing photography has both a number of advantages and associated problems, as follows:

Advantages

  • The photos provide a visual assessment of relatively remote areas
  • The relatively low cost of obtaining photography
  • Photos can be found from the present day back to the 1940's, providing insights into habitat change

Disadvantages

  • Coverage of the photography is patchy - both through time and space
  • Maximum water depths for bottom visibility are often <10 m
  • The photography was almost certainly acquired to map the land - not the water - and many areas of shallow water are missed
  • Sun glint and waves can render an image virtually useless
  • The time of image acquisition is often not optimal for water clarity
  • Mapping of highly dynamic habitats is complicated because the extent and condition of the habitat can change rapidly with the seasons and even storm events. The date of acquisition of the pre-existing images often is not consistent with important habitat fluctuations

Planned acquisition of aerial imagery

A postgraduate research project associated with the SEAMAP project is looking into optimising the capture of remote sensing imagery for both the mapping and monitoring of habitats in shallow marine waters, particularly seagrass beds.

The successful capture of images clearly showing the sea floor is a challenging remote sensing problem mostly due to the added complications created by the presence of two mediums - air and water - through which the light must travel. Water can be thought of acting like a thick atmosphere, distorting and diffusing the light ray paths both within the water body and as it crosses the air\water boundary. Added to that are the high levels of absorption and deflection of light beams caused by particles suspended in the water such as plankton and sediment. Many of Tasmania's waters are also rich in the dark tea-coloured tannic river waters further limiting the penetration of light into the water column.

In spite of the limiting factors, enough promise is indicated by the pre-existing images to warrant further development of remote sensing techniques. The major research questions are:

  • What are the water clarity and depth limitation on seagrass visibility in high resolution aerial photography?
  • Can remote sensing detect the deeper boundary edge necessary for determining bed extent?
  • Can seagrass extent and patchiness be measured in a way that allows comparisons both of the same bed through time (ie change detection) and of different beds?
  • To what extent can measurements of seagrass biomass, epiphyte and associated algal biomass be made efficiently, reliably and repeatedly using a combination of remote sensing and field sampling?

The research is mostly based on high resolution digital aerial photography. The choice of sensor and platform is driven by both economic considerations and the high degree of flexibility they give over image acquisition.

Progress to date indicates that very clear imagery in Tasmanian is possible. Some factors to take into account for successful image acquisition are:

  • Best results are obtained in the early morning when wind speeds are low and while the sun angles are low enough to bounce reflections under the sensor. Care needs to be taken with cloud cover to ensure surface reflections from diffuse light are minimised. A paper outlining a simplified method of accurately predicting the amount of sun glitter in an image using time of day, location, wind speed and camera details is currently in review.
  • The most suitable time of year is March to May when the higher plankton levels of summer are clearing and turbid river water inputs are low, yet the seagrass biomass is still high.
  • Ground control is an important issue for rectification but is often difficult in the marine environment due to the lack control points. Careful planning of the location for camera exposure stations can maximise the use of the coastline, if available. Until full camera inertial movement systems are readily available, temporary floating ground control markers can be deployed to assist with image rectification and orthophoto creation.
  • Orthophoto production over water is complicated by refraction. A method to remove the displacement in the image created by both relief and refraction has been developed and implemented in an ArcInfo environment. It relies on a monoscopic solution and requires a DEM of the area.

The acquired imagery is to be used in developing techniques for monitoring and mapping seagrass extent, patchiness, biomass and epiphytic loading. To this end, we have combined the image acquisition with contemporaneous boat-based collection of biomass and epiphyte loading data using diver an

d video sampling methods. Analysis methods will include:

  • The application of spatial metrics at the landscape scale,
  • Estimation of biological parameters such as biomass and epiphytic loading,
  • Fuzzy classification of seagrass and image segmentation techniques.

Satellite Imagery

Rapid Eye

We are currently invesitgating the use of Rapid Eye satellite imagery for marine habitat mapping.

The aim is to use high resolution multispectral satellite imagery from RapidEye to quantify the distribution of Macrocystis pyrifera in eastern and southern Tasmania. There are three main aspects to this project. The first involves reviewing the methodology to extract a signal for M. pyrifera from satellite multispectral remotely sensed data. Based on this information, a methodology will be refined to assess the current distribution of M.pyrifera based on data from the RapidEye satellite. Lastly, an assessment of the conservation status of M. pyrifera habitatas a distinct coastal community type in Tasmania will be carried out. The research questions that will be focus of this study are:

1)    What is the most efficient method to extract the M. pyrifera signal from the RapidEye data, and how can this be adequately ground truthed?

2)    What is the current distribution of M. pyrifera around eastern and southern Tasmania?

3)    What are the implications of the current distribution and trend change in M. pyrifera for the current exercise assessing the conservation status of M. pyrifera habitatas a distinct coastal community type in Tasmania?

To see the Rapid Eye Gallery follow this link.