What is ProRaster Scientific?

ProRaster Scientific is software for rendering, processing, and analysing Earth Observation imagery. The advanced virtual raster processing pipeline technology minimises time, costs, and computation resources, whilst maximising efficiency, performance, and data quality.

ProRaster Scientific delivers three major capabilities –

How does Multispectral Imagery Processing Work?

ProRaster Scientific has automated support for all Landsat imagery and Sentinel2 imagery, and manual support for other platforms. It uses the power of virtual rasters to eliminate data duplication and maximise processing efficiency, saving you time and money.

To get started with multispectral imagery in ProRaster Scientific, you will follow a processing pipeline like this –

1. Download scenes or tiles from your data provider.

The Earth Observing Satellites acquire data in swaths called a “scene”. Each scene covers a static footprint on the Earth and the satellite will typically revisit that scene regularly and reacquire it many times over the life of the bird. This data is processed and then distributed to users as a package containing rasters for each spectral band and other information. This is a scene, and it is your starting point. A tile is a rectangular subset of a scene.

You can download Landsat scenes for free from USGS Earth Explorer –
https://earthexplorer.usgs.gov/

You can download Sentinel 2 tiles for free from the Copernicus Browser accessed via the Copernicus Data Space Ecosystem –

https://dataspace.copernicus.eu/

2. Unpack and store your scenes on your local drive or network.

The scene/tile will be delivered as a compressed archive. You will need to decompress (unpack) this archive and store the data package on your local drive or network. How you organise this is up to you.

The data package generally consists of multiple rasters, each of which contains observational data over a spectral frequency range, as well as other ancillary rasters. It can be difficult to know what files correspond to what spectral bands, but ProRaster takes care of those details for you. 

3. Create a multispectral scene database and add the scenes to the database.

In ProRaster, add each scene/tile you download to your scene database. The database is a small file that links to the original raster data – it does not contain a copy of the raster data. ProRaster will generate a “scene assembly virtual raster” for each new scene/tile and store it alongside the database. You will access all data from the scene via this assembly MVR from now on.

The scene assembly MVR will apply all the corrections required to translate the raw digital numbers supplied to reflectance.

4. Render the scenes.

Take a look at your new scene! Open the scene database browser and select the scene. Right-click and from the pop-up context menu, select “Render Scene”. Choose a band combination to highlight specific attributes of the Earth. A rendering algorithm will be created for your scene in the main ProRaster interface, and the scene will be displayed in a map.

5. Create a multispectral product database.

In ProRaster we distinguish between a basic scene and a multispectral data product. A product starts life as one or more scenes, to which you can apply processing operations in a waterfall cascade pattern. Before you can create a product, you need to create a database to store it in.

6. Browse for and select scenes from which to create a multispectral product.

Open a scene database browser and select a scene. Right-click and from the pop-up context menu, select “Create scene product”. A new product will be added to your product database and the new product will be opened in the product editor dialog.

7. Edit the product and apply processing operations.

You can apply a growing list of processing operations to the scene in your product including – Mask, Clip, Index, Reproject, Calculator, Transform, and Difference. You can generate a rendering algorithm, compute statistics, and export the data to a raster. Define these operations in any order you wish to build a waterfall cascade processing pattern. Thereafter, when you change any operation, all the downstream operations will get updated.

As your level of expertise improves, you will go on to combine products into more advanced products and combine rendering algorithms into more advanced visualisations. Soon, you will be using the “Create Product Sequence Wizard” to build advanced temporal sequence (multidimensional) products.

How does Raster Processing work?

Traditional raster processing imports a source raster and applies a processing operation to it, then exports an output raster. This process consumes time, computing resources and storage, and you can’t check the result of your processing operation until the output has been generated.

ProRaster Scientific takes a different approach. It uses virtual raster technology to apply processing operations in real-time, at any scale, to rasters of any size. It takes no time, uses no resources, and the results can be rendered immediately allowing you to see if your processing operation has acted appropriately.

As an example, consider a very large gigapixel scale raster. If you were to clip this raster to a polygon and output a new raster it might takes minutes or hours to execute the processing. But in ProRaster we simply create a virtual raster (which takes no time and uses no resources) then immediately render it. The clipping occurs in real-time as you render the raster.

Furthermore, the virtual rasters we create as output from each operation can be used as input to another operation, so you can build chains of processing operations. The multispectral product editor leverages this technology.

To take your data into other platforms, crystallise your virtual raster at any time by exporting it to a standard raster format like GeoTIFF or MRR.

How does Raster Rendering work?

All raster data is rendered in ProRaster by building a rendering algorithm. A rendering algorithm can be saved to disk as an MRD file. If you take a look at one, you will find it just contains standard XML. 

The rendering engine in ProRaster takes a rendering algorithm and, following the instructions therein, converts the raster data to colored pixels. It uses multithreading and hardware graphics acceleration to get those pixels into your map as quickly as possible for immersive, fluid rendering.

The rendering engine is laser-focused on outputting the highest quality imagery at all times. You do not have to worry about raster cell size, cell alignments, coordinate systems, or data interpolation. Give it any raster data, in any combination, with any properties, and it will give you pixel-perfect imagery every time.

Here is a high level overview of rendering algorithm capabilities – 

• Create, load, and edit rendering algorithms and save to MRD format.
• Combine multiple Image, LUT Color, or RGB Color layers.
• Image layers combine RGB(A) or Greyscale Imagery with Intensity modulation, Opacity modulation, and Pixel Masking components.
• Color LUT layers combine a Color component (which maps data values into a color lookup table) with Intensity modulation, Opacity modulation, and Pixel Masking components.
• RGB Color layers combine Red, Green, and Blue components (which map data values to color) with Pan-sharpening, Intensity modulation, Opacity modulation, and Pixel Masking components.
• Manage your Color Table, Data Transform, Data Conditioning Filter, and Raster Source global resources through comprehensive editors.

Basic Training

The following videos are best viewed on YouTube and will help you learn how to use ProRaser Scientific. Don’t forget to “like and subscribe” and if you leave a comment or question, we will be able to respond to it.

ProRaster Scientific 2.4.01 – 3D Maps Deep Dive Tutorial

ProRaster Scientific version 2.4.01 supports 3D maps in which you can display one or more 3D surfaces draped with imagery, directly from your 2D rendering algorithm. Take a deep dive into 3D surface rendering and learn how to use this feature in ProRaster Scientific.

Classify Raster Processing Operation

See how to use the new Classify (raster) processing operation in ProRaster Scientific 2.2.09.

New User Interface Features

I take you on a walk-around of the new user interface features in ProRaster Scientific.

ProRaster Scientific builds on – and extends – ProRaster Premium. This video takes a quick look at the new user interface elements in ProRaster Scientific.

Downloading Landsat Scenes

I take you through the basics of downloading a Landsat scene from Earth Explorer and importing it into ProRaster Scientific.

Rendering Satellite Imagery

See all the different ways that you can render a multispectral satellite imagery scene in ProRaster Scientific.

I build algorithms manually, via the scene database browser, and via the product editor. There are many ways to render multispectral data in ProRaster Scientific and it is important to be familiar with the most efficient and powerful mechanisms.

The Multispectral Scene Database

Learn how to create and maintain multispectral satellite scene databases in ProRaster Scientific. See demonstrations of how to import scenes and find out what the scene database does, and where it is stored.

Using the Scene Database Browser

Learn how to use the multispectral satellite scene browser to query your scene database and find the scenes you want.

See how to create products from the browser including scenes, mosaics, collated scenes and mosaics, and scene sequences.

The Multispectral Product Database

Learn how to create and maintain multispectral satellite product databases in ProRaster Scientific. Find out what the product database does, and where it is stored. See how to make products from scenes and how to edit products.

The Multispectral Product Editor

Learn how to build virtual processing cascades in the multispectral satellite imagery product editor in ProRaster Scientific. See an overview of the operations available and learn how to control operation execution.

The Create Product Sequence Wizard

Learn how to turbocharge your productivity with the powerful new “Create Product Sequence” wizard in ProRaster Scientific version 2.2.

Working with Spot7 Imagery

Learn how to manually import a Spot7 multispectral satellite scene into ProRaster Scientific. With high-resolution Blue, Green, Red, and Near Infrared spectral bands, Spot7 data provide ample opportunity for vegetation analysis.

Working with GeoEye1 Imagery

Learn how to manually import a GeoEye1 multispectral satellite scene into ProRaster Scientific. With high-resolution Blue, Green, Red, and Near Infrared spectral bands and with a very high-resolution panchromatic band, GeoEye1 data provide ample opportunity for vegetation analysis.

Before watching this video, watch the Spot7 video which covers the procedure in more detail.

Introducing the Processing Menu

Explore the new Raster Processing operations in ProRaster Scientific. The operations generate virtual raster outputs and so operations on huge rasters can be visualized immediately, without executing any processing!

Worked Examples

The following videos are best viewed on YouTube and show how to use ProRaster Scientific in worked examples using real data and analysis.

Quantifying vegetation health at Macquarie Island before and after pest species eradication

I use freely available Landsat Earth Observation multispectral imagery and ProRaster Scientific to quantify vegetation health at Macquarie Island before and after a long-term pest species eradication program.

Located halfway between New Zealand and Antarctica, Macquarie Island is a cold and windswept haven for Southern Ocean sea birds, penguins, and marine mammals. A program has been conducted by Australian governments to eradicate pest species on the island including rabbits, rats, and mice. These pest species were degrading the island vegetation and impacting the breeding behaviour of resident and visiting wildlife.

I acquire Landsat 7 scenes from 1999-2002 and Landsat 8 & 9 scenes from 2022-2024 to compare NDVI (vegetation index) statistics before the eradication program and after it. The data show a stark improvement that testifies to the success of the program.

Preparing GLO-30 DEM Data in ProRaster Scientific

GLO-30 is a one-arcsecond (30 metre) resolution global DEM raster that can be acquired for free from the ESA via Copernicus. You can download 1×1 degree tiles in TIFF format as packages for countries and regions.

In the first video, I show you how to take the tile collection for a country and work with it in ProRaster Scientific. Firstly, I create raster sources for the DEM files and the WBM water mask files. Secondly, I use the “Create Raster Mask” operation to create a mask from the WBM files that retains pixels classified as Land, Rivers, and Lakes but removes Ocean. Thirdly, I use the “Clip to Raster” operation to clip the DEM to the Ocean mask. Finally, I export this virtual raster (MVR) to a physical raster (MRR). From almost 2.5 GB of original source data, I create a high-quality DEM raster in MRR format that only uses 26 MB – a 100x improvement in storage efficiency.

In the second video, I show you how to prepare the mask rasters using the Classify processing operation. These include the Height Error Mask, the Water Body Mask, the Filling Mask, and the Edit Data Mask.

Maping the 2022 Indus Valley Flood, Pakistan

I compute the area of Pakistan that was flood affected in the 2022 flood event in the Indus Valley. This was stimulated by a three-way comparison where experts used Google Earth Engine, Microsoft Planetary Computer, and QGIS to make this computation. As a result of this analysis, I added some new compositing features to ProRaster Scientific. In the second video, I use these new techniques to come to a final estimate of 27,772 square kilometres. 

GEBCO 2023 Gridded Bathymetry Data in ProRaster Scientific

Each year GEBCO releases an updated bathymetry raster for the world at 15 arc second resolution. I download the eight TIFF files and show you how to display them in ProRaster Scientific. My first approach is to use the Raster Source Editor to create a raster source for the TIFF files. Then, I use a Join operation to create a virtual raster for this raster source. Finally, I export the virtual raster to an MRR format raster. I take a look at the bathymetry off the northwest coast of the Kimberly in Australia, then lower the sea level by 91 metres to reveal the paleo-topography of this region.

Coorabulka Revisited – NDVI Analysis

Taking advantage of new productivity tools in ProRaster Scientific version 2.2.03, I repeat the Coorabulka NDVI analysis in real time in a single video! We start with 552 Landsat 8 and Landsat 9 scenes, Collection 2 Level 1. I create a scene database, then use the Create Product Sequence Wizard to generate a Collated Mosaic Sequence product over more than 10 years. I then compute NDVI, clipping the data using a calculator expression, and clipping the raster spatially to the Coorabulka Station polygon. I compute statistics and then convert NDVI (Normalised Difference Vegetation Index) from a vegetation index to a proxy measure of “Feed On Offer” using a non-linear data transformation.

ProRaster Scientific NDVI Analysis – Coorabulka Station – Part 1 of 6

In this six-part series, I use ProRaster Scientific to perform a vegetation change analysis at Coorabulka Station in western Queensland over an entire year from July 2022 to July 2023. Coorabulka is located in the “channel country”, where infrequent flooding events can transform the station’s vegetation and carrying capacity.

In Part 1, I take a look at Coorabulka and the hardware I use to perform the processing, then build the multispectral satellite imagery scene database in ProRaster Scientific.

ProRaster Scientific NDVI Analysis – Coorabulka Station – Part 2 of 6

In this six-part series, I use ProRaster Scientific to perform a vegetation change analysis at Coorabulka Station in western Queensland over an entire year from July 2022 to July 2023. Coorabulka is located in the “channel country”, where infrequent flooding events can transform the station’s vegetation and carrying capacity.

In Part 2, I show you how to build mosaics from satellite scenes and then discuss the advantages of using collation to improve your data quality in a mosaic. We put these mosaic products together into a mosaic sequence that adds a temporal dimension to our raster dataset.

ProRaster Scientific NDVI Analysis – Coorabulka Station – Part 3 of 6

In this six-part series, I use ProRaster Scientific to perform a vegetation change analysis at Coorabulka Station in western Queensland over an entire year from July 2022 to July 2023. Coorabulka is located in the “channel country”, where infrequent flooding events can transform the station’s vegetation and carrying capacity.

In Part 3, we go deeper into building high-quality Collated Mosaic Sequences that draw in data from multiple scene revisits both before and after the target visitation time. We can then render the entire data sequence using an enhanced Natural Color band ratio.

ProRaster Scientific NDVI Analysis – Coorabulka Station – Part 4 of 6

In this six-part series, I use ProRaster Scientific to perform a vegetation change analysis at Coorabulka Station in western Queensland over an entire year from July 2022 to July 2023. Coorabulka is located in the “channel country”, where infrequent flooding events can transform the station’s vegetation and carrying capacity.

In Part 4, we take the collated mosaic sequence and apply a cascade of processing operations to it. This includes computing NDVI, clipping the data to a polygon of Coorabulka, applying raster calculator operations, and computing statistics. I go into detail for each of these operations, as well as give you some tips on how to work with the Product Editor.

ProRaster Scientific NDVI Analysis – Coorabulka Station – Part 5 of 6

In this six-part series, I use ProRaster Scientific to perform a vegetation change analysis at Coorabulka Station in western Queensland over an entire year from July 2022 to July 2023. Coorabulka is located in the “channel country”, where infrequent flooding events can transform the station’s vegetation and carrying capacity.

In Part 5, we review our results as temporal imagery. We look at the mosaic sequence, the collated mosaic sequence, and the NDVI sequence for Coorabulka.

ProRaster Scientific NDVI Analysis – Coorabulka Station – Part 6 of 6

In this six-part series, I use ProRaster Scientific to perform a vegetation change analysis at Coorabulka Station in western Queensland over an entire year from July 2022 to July 2023. Coorabulka is located in the “channel country”, where infrequent flooding events can transform the station’s vegetation and carrying capacity.

In Part 6, I examine the temporal statistics in Excel and draw conclusions. I graph all the statistical measures and explain what they are telling us about the change in vegetation at the station. I create a vegetation biomass proxy measure and discuss the reasoning behind such a measure, and how it can be related to stocking rates and station carrying capacity.

ProRaster Scientific – Wooroloo Fire Scar Analysis

The Wooroloo fire burned through the foothills east of Perth, Western Australia, from February 1 to 6, 2021.

Using Landsat 8 and 9 surface reflectance data, I build a scene sequence product consisting of six events, and a processing operation cascade. This includes computing the BAI (Burned Area Index) spectral index, applying a calculator operation, clipping to a polygon, and computing detailed statistics.

ProRaster Scientific – Tjoritja Fire Scar Analysis

The Tjoritja (West MacDonnell) National Park is located west of Alice Springs in Central Australia. In March 2023 it suffered extensive damage from an uncontrolled bushfire.

I use Landsat 8 & 9 surface reflectance imagery to analyse the fire scar. We will track the fire scar over time and use the NBRT1 index to map the extent. In this video, you will see an example of how to build branched processing chains in the multispectral satellite imagery Product Editor.

Pre-First Release Examples

The following videos are best viewed on YouTube and were made during the development of ProRaster Scientific to guide and inform the software development.

Mapping 50 years of Bauxite Mining in Western Australian Jarrah Forests with Landsat

I map the extent of bauxite mining in the Jarrah forests southeast of Perth in Western Australia from 1974 through to February 2022 using Landsat satellite multispectral imagery.

Prompted by the article “Undermined”, by Victoria Laurie in the August 13, 2022 edition of the Weekend Australian, I investigate how 50 years of Landsat data can be used to map the extent of mining operations over time.

I use ProRaster Scientific to build a time series of Landsat scenes from 1974 to 2022. I use Surface Reflectance where available. In earlier scenes, this drops back to top-of-atmosphere reflectance, and in the earliest scenes I use MSS instrument data.

I manually polygonise the mining operations extents for each scene and then combine these polygons in MapInfo Pro into a single complex polygons, for which I can compute the area.

I compute the NDVI index for the 2022 surface reflectance scene to see whether this index can be used to differentiate between undisturbed forest and rehabilitated areas but conclude that it cannot be used.

The Making of “Mapping 50 years of Bauxite Mining in Western Australian Jarrah Forests with Landsat”

In this video, I describe the technology in ProRaster Scientific that was used to make the Bauxite mapping video.

Sneak Peek: Pansharpening in ProRaster Scientific

Pansharpening is a process that improves the apparent resolution of satellite multispectral data at render time by integrating a high-resolution panchromatic band into the RGB color data acquired from three spectral bands. In this video, I show you how to perform panchromatic sharpening of a satellite scene. In ProRaster, pansharpening is performed on the fly so there is no preparation or processing required. You can control the brightness of the imagery via the data transform for the pan component.

Sneak Peek: Importing Landsat and Sentinel 2 scenes into ProRaster Scientific

ProRaster Scientific is under development and will focus on rendering and processing satellite multispectral data.

In the video below, I give you a sneak peek into the import process. This will analyse a satellite scene that you have acquired (for example, from the USGS) and generates a virtual raster that contains all the data for the scene, as well as value-added products.

ProRaster Scientific can import Landsat and Sentinel 2 scenes.

Landsat Preprocessing for ProRaster Scientific

The video below discusses the ins and outs of pre-processing Landsat data. ProRaster Scientific will help with this process and it is complicated and confusing to come to grips with. I hope this video helps explain the details.