Geoscience Australia has a variety of satellite data processing options for optical data through Landsat and other satellites and sensors. In depth technical information is available by contacting Earth Observation Client Services.
The data collected and transmitted to Earth from an Earth observation satellite is in a form unsuitable for use by customers. To enable visualisation, we process this raw data in varying degrees to generate products which cover a range of processing levels. Raw data is generally processes only after you have placed an order for a specific product. This is due to the large quantities of data received from a satellite and the wide range of product options available. Storing data in its raw form is more efficient in terms of storage and facilitates improved processing of data as processing systems improve over time.
The processing of satellite data involves two basic types of corrections:
The sensing instruments on Earth observation satellites contain a multitude of detectors for each band. For example, the Landsat 5 TM instrument has 16 detectors for each band. The detectors convert the observed radiance (the reflected energy from the Earth's surface) into an electronic signal. However, not all detectors will do this in an identical manner. This may lead to striping across the image.
The most common method for correcting the striping is based on the assumption that the statistics of the detector signals (brightness values) will be the same for all detectors. The success of this method will rely on a good sample of imagery being used to generate the statistics of the detector brightness values
A reference detector for each band is then chosen and the brightness values for the other detectors adjusted to match the brightness values of the reference detectors.
The radiometric correction of imagery described above is often referred to as relative radiometric calibration. This is the type of calibration applied to Geoscience Australia's optical data.
Adjacent digital scenes of the same acquisition date will exhibit similar brightness values for the same ground feature. However, if the acquisition dates are different, there is more likely to be a significant difference between the brightness values. These differences in brightness values may be the result of changes in atmospheric conditions, sun angles and reflectance characteristics of ground surface features between the dates of acquisition.
The absolute radiometric calibration of data involves the conversion of the pixel brightness values into units of spectral radiance. This type of calibration is not done for any of our products, and, if required, must be performed by the user.
The raw data acquired by the sensor is transmitted and recorded as a sequential stream of pixels with no image structure. The geometric correction process restores this data to a two dimensional image using sets of corrections to enable it to be interpreted as a spatial representation of the Earth's surface.
The corrections are based on mathematical models which attempt to account for the various physical factors involved in the imaging process. The models are:
Sensor model which accounts for:
Satellite model which accounts for:
Earth model which accounts for:
Geoscience Australia's products may be generated as either path oriented or map oriented.
Path oriented
For path oriented products the imagery (rows and columns of pixels) is still aligned with the direction of the satellite orbit, which is about 10 degrees east of north for optical sensors on a descending swath over Australia. These products are not suitable for direct input into a GIS system and must first be rotated by you to align with a map grid.
Map oriented
In map oriented products, the imagery (rows and columns of pixels) has been transformed to a map grid and the rows and columns of image pixels are resampled to align with map-grid north. In this resampling process, the resulting pixel size of products may change from the original pixel size.
The level of geometric correction applied to a satellite image also can be divided into three other categories labelled systematic, precision and orthocorrected.
Systematic products
Systematic products are corrected by determining the geographic location of the image using a satellite model based on information about spacecraft position and attitude. This information is transmitted with the imagery, or provided by the satellite operator at the time of, or following image acquisition. It allows the imagery to be located to a limited absolute accuracy (ie. its position on the Earth).
Precision products
If precision products are offered, the satellite model is refined by the use of Ground Control Points (GCPs). These are points which can be accurately identified in both the remotely sensed image and a known reference source (eg. topographic map, GPS points) and then used to appropriately rotate the image to the known GCPs. The accuracy of these products depends on the accuracy of the GCPs used.
Orthocorrected products
Orthocorrected products are precision products where the effects of relief displacement are corrected by using constraints from a Digital Elevation Model (DEM). Fundamentally this corrects for the effect of the received signal being at a lower angle (or closer from vertical) to the satellite, rather than actually being elevated and at a higher angle (or further away). The accuracy of these products depends on the accuracy of the GCPs and on the resolution of the DEM.
If precision or orthocorrected products are required over non-Australian territories, customers must supply the appropriate maps or GCPs and/or DEM.
As part of the image correction process, brightness values must be derived for each pixel in the image output. The value is interpolated from the uncorrected satellite image. This interpolation exercise is known as resampling.
More information is available at Geometric Correction and Resampling of Data.
Geoscience Australia has three main processing options:
We also have two additional non standard options:
If available, we recommends our orthocorrected image for most applications where it can be readily and accurately related to a geographic location and used with other geographic information.
Please check with Geoscience Australia , or your preferred distributor for the most appropriate processing option for your application.
| Main Options | Description |
|---|---|
| Orthocorrected Image | Map oriented with systematic geometric corrections refined with the use of GCPs and the best DEM available to Geoscience Australia (which is equal to or better than the Geoscience Australia 9 second National DEM V2). Two dimensional resampling to align with a map grid. This is the most map-accurate product, suitable for input into a GIS. |
| Map Oriented Image | Map oriented with systematic radiometric and geometric corrections applied to the data. Two dimensional resampling to align with a map grid. |
| Path Image | Satellite path oriented with systematic radiometric and geometric corrections applied to the data. Two dimensional resampling to fit a specific Earth datum and map grid. |
| Non-standard Options | |
| Raw Image | Radiometrically corrected but no geometric corrections. |
| Raw Data | No radiometric or geometric corrections. Not recommended for customer use |
More details on the processing of optical data from specific sensors can be found at:
Geoscience Australia products are extracted from a continuous swath of data collected by the orbiting satellite. Framing refers to the selection of a smaller part of the swath or standard size scene image. Products can be framed from the swath in a variety of ways. Geoscience Australia and its distributors are experienced in selecting the most appropriate framing for your requirements. In some cases we will request a map of your area of interest to ensure correct framing.
The Earth Observation Catalogue also provides the ability to view the framing of data products within the satellite swath. It also allows for the export of mid/mif files to facilitate viewing of quicklook images in MapInfo software.